{"repository":"Qwen/Qwen3.8-Flash-Next","revision":"de4b8e4d43b917e7706784d8bb445c9af86a3540","method":"Safetensors header shapes only; bounded HTTP Range requests; no tensor payload materialized","shards":131,"tensor_count":1658,"index_sha256":"99e815241ef03325536b0aaa4441deea45174c17fae31e10f0bb456410c590de","config_sha256":"889658f2508e8c61d409b02e70e0d78d8d4452ec65aaafbe129805d213d2e74b","dtype_elements":{"BF16":179999981424,"I64":35},"weight_bytes":359999963128,"groups":{"gated_residual":640624640,"gdn":2086510464,"vision":448931056,"routed_experts":120795955200,"router":62914560,"shared_expert":236052480,"integer_buffers":35,"lexical_injection":32839680,"lexical_table":51200245760,"mtp":2607150848,"qsa":617358336,"input_embedding":635699200,"output_head":635699200},"tensors":{"model.language_model.hyper_connection_mixer.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00001-of-00131.safetensors"},"model.language_model.hyper_connection_mixer.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00001-of-00131.safetensors"},"model.language_model.hyper_connection_mixer.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.0.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.1.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.10.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.11.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.12.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.13.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.14.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.15.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.16.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.17.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.18.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.19.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.2.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.20.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.21.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.22.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.23.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.24.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.25.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.26.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.3.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.4.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.5.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.6.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.7.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.8.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.attn.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.attn.proj.weight":{"shape":[1152,1152],"dtype":"BF16","numel":1327104,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.attn.qkv.bias":{"shape":[3456],"dtype":"BF16","numel":3456,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.attn.qkv.weight":{"shape":[3456,1152],"dtype":"BF16","numel":3981312,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.mlp.linear_fc1.bias":{"shape":[4304],"dtype":"BF16","numel":4304,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.mlp.linear_fc1.weight":{"shape":[4304,1152],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.mlp.linear_fc2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.mlp.linear_fc2.weight":{"shape":[1152,4304],"dtype":"BF16","numel":4958208,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.norm1.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.norm1.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.norm2.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.blocks.9.norm2.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.merger.linear_fc1.bias":{"shape":[4608],"dtype":"BF16","numel":4608,"shard":"model-00001-of-00131.safetensors"},"model.visual.merger.linear_fc1.weight":{"shape":[4608,4608],"dtype":"BF16","numel":21233664,"shard":"model-00001-of-00131.safetensors"},"model.visual.merger.linear_fc2.bias":{"shape":[2560],"dtype":"BF16","numel":2560,"shard":"model-00001-of-00131.safetensors"},"model.visual.merger.linear_fc2.weight":{"shape":[2560,4608],"dtype":"BF16","numel":11796480,"shard":"model-00001-of-00131.safetensors"},"model.visual.merger.norm.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.merger.norm.weight":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.patch_embed.proj.bias":{"shape":[1152],"dtype":"BF16","numel":1152,"shard":"model-00001-of-00131.safetensors"},"model.visual.patch_embed.proj.weight":{"shape":[1152,3,2,16,16],"dtype":"BF16","numel":1769472,"shard":"model-00001-of-00131.safetensors"},"model.visual.pos_embed.weight":{"shape":[2304,1152],"dtype":"BF16","numel":2654208,"shard":"model-00001-of-00131.safetensors"},"model.language_model.layers.0.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00002-of-00131.safetensors"},"model.language_model.layers.0.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.0.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.0.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.0.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.0.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.0.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.0.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.0.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.0.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.0.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00003-of-00131.safetensors"},"model.language_model.layers.1.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00004-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.layer_multipliers":{"shape":[3],"dtype":"I64","numel":3,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.ple.key_proj.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.ple.norm_conv.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.ple.norm_key.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.ple.norm_query.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_0.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_1.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00005-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_2.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00006-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_3.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00006-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_4.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00006-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_5.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00006-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_6.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00007-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_7.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00007-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_8.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00007-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_9.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00007-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_10.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00008-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_11.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00008-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_12.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00008-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_13.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00008-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_14.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00009-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_15.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00009-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_16.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00009-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_17.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00009-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_18.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00010-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_19.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00010-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_20.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00010-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_21.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00010-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_22.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00011-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_23.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00011-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_24.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00011-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_25.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00011-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_26.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00012-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_27.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00012-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_28.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00012-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_29.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00012-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_30.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00013-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_31.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00013-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_32.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00013-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_33.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00013-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_34.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00014-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_35.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00014-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_36.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00014-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_37.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00014-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_38.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00015-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_39.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00015-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_40.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00015-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_41.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00015-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_42.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00016-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_43.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00016-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_44.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00016-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_45.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00016-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_46.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00017-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_47.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00017-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_48.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00017-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_49.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00017-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_50.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00018-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_51.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00018-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_52.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00018-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_53.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00018-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_54.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00019-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_55.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00019-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_56.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00019-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_57.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00019-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_58.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00020-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_59.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00020-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_60.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00020-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_61.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00020-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_62.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00021-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_63.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00021-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_64.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00021-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_65.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00021-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_66.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00022-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_67.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00022-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_68.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00022-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_69.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00022-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_70.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00023-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_71.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00023-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_72.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00023-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_73.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00023-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_74.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00024-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_75.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00024-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_76.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00024-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_77.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00024-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_78.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00025-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_79.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00025-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_80.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00025-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_81.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00025-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_82.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00026-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_83.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00026-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_84.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00026-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_85.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00026-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_86.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00027-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_87.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00027-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_88.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00027-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_89.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00027-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_90.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00028-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_91.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00028-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_92.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00028-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_93.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00028-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_94.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00029-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_95.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00029-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_96.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00029-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_97.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00029-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_100.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00030-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_101.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00030-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_98.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00030-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_99.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00030-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_102.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00031-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_103.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00031-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_104.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00031-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_105.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00031-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_106.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00032-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_107.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00032-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_108.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00032-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_109.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00032-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_110.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00033-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_111.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00033-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_112.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00033-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_113.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00033-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_114.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00034-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_115.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00034-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_116.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00034-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_117.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00034-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_118.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00035-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_119.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00035-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_120.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00035-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_121.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00035-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_122.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00036-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_123.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00036-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_124.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00036-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_125.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00036-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_heads_offsets":{"shape":[16],"dtype":"I64","numel":16,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_heads_vocab_sizes":{"shape":[16],"dtype":"I64","numel":16,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.1.ple.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_126.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.1.ple.ple_embedding.ngram_embedding.shard_127.weight":{"shape":[2500012,160],"dtype":"BF16","numel":400001920,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.1.ple.value_proj.weight":{"shape":[2560,2560],"dtype":"BF16","numel":6553600,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00037-of-00131.safetensors"},"mtp.hyper_connection_mixer.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00037-of-00131.safetensors"},"model.language_model.layers.10.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00038-of-00131.safetensors"},"model.language_model.layers.10.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.10.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.10.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.10.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.10.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.10.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.10.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.10.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.10.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.10.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.11.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.11.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.11.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.11.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00039-of-00131.safetensors"},"model.language_model.layers.11.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00040-of-00131.safetensors"},"model.language_model.layers.11.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.11.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00041-of-00131.safetensors"},"mtp.hyper_connection_mixer.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00041-of-00131.safetensors"},"model.language_model.layers.12.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00042-of-00131.safetensors"},"model.language_model.layers.12.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.12.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.12.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.12.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.12.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.12.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.12.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.12.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.12.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.12.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00043-of-00131.safetensors"},"mtp.hyper_connection_mixer.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00043-of-00131.safetensors"},"model.language_model.layers.13.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00044-of-00131.safetensors"},"model.language_model.layers.13.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.13.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.13.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.13.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.13.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.13.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.13.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.13.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.13.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.13.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00045-of-00131.safetensors"},"model.language_model.layers.14.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00046-of-00131.safetensors"},"model.language_model.layers.14.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.14.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.14.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.14.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.14.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.14.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.14.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.14.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.14.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.14.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.15.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.15.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.15.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.15.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00047-of-00131.safetensors"},"mtp.layers.0.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00047-of-00131.safetensors"},"model.language_model.layers.15.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00048-of-00131.safetensors"},"model.language_model.layers.15.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.15.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00049-of-00131.safetensors"},"mtp.layers.0.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00049-of-00131.safetensors"},"model.language_model.layers.16.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00050-of-00131.safetensors"},"model.language_model.layers.16.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.16.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.16.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.16.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.16.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.16.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.16.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.16.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.16.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.16.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00051-of-00131.safetensors"},"model.language_model.layers.17.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00052-of-00131.safetensors"},"model.language_model.layers.17.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.17.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.17.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.17.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.17.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.17.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.17.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.17.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.17.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.17.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00053-of-00131.safetensors"},"mtp.layers.0.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00053-of-00131.safetensors"},"model.language_model.layers.18.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00054-of-00131.safetensors"},"model.language_model.layers.18.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.18.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.18.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.18.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.18.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.18.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.18.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.18.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.18.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.18.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.19.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.19.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.19.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.19.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00055-of-00131.safetensors"},"model.language_model.layers.19.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00056-of-00131.safetensors"},"model.language_model.layers.19.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.19.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00057-of-00131.safetensors"},"mtp.layers.0.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00057-of-00131.safetensors"},"model.language_model.layers.2.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00058-of-00131.safetensors"},"model.language_model.layers.2.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.2.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.2.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.2.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.2.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.2.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.2.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.2.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.2.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.2.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.20.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.20.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.20.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.20.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.20.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00059-of-00131.safetensors"},"model.language_model.layers.20.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00060-of-00131.safetensors"},"model.language_model.layers.20.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00060-of-00131.safetensors"},"model.language_model.layers.20.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00060-of-00131.safetensors"},"model.language_model.layers.20.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00060-of-00131.safetensors"},"model.language_model.layers.20.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00060-of-00131.safetensors"},"model.language_model.layers.20.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00060-of-00131.safetensors"},"model.language_model.layers.20.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00060-of-00131.safetensors"},"model.language_model.layers.20.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00060-of-00131.safetensors"},"mtp.layers.0.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00060-of-00131.safetensors"},"model.language_model.layers.20.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00061-of-00131.safetensors"},"model.language_model.layers.20.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.20.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.20.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.20.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.20.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.20.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.20.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.20.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.20.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.20.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00062-of-00131.safetensors"},"mtp.layers.0.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00062-of-00131.safetensors"},"model.language_model.layers.21.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00063-of-00131.safetensors"},"model.language_model.layers.21.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.21.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.21.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.21.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.21.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.21.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.21.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.21.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.21.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.21.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00064-of-00131.safetensors"},"model.language_model.layers.22.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00065-of-00131.safetensors"},"model.language_model.layers.22.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.22.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.22.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.22.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.22.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.22.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.22.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.22.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.22.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.22.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.23.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.23.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.23.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.23.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00066-of-00131.safetensors"},"mtp.layers.0.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00066-of-00131.safetensors"},"model.language_model.layers.23.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00067-of-00131.safetensors"},"model.language_model.layers.23.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.23.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00068-of-00131.safetensors"},"model.language_model.layers.24.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00069-of-00131.safetensors"},"model.language_model.layers.24.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.24.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.24.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.24.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.24.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.24.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.24.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.24.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.24.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.24.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00070-of-00131.safetensors"},"mtp.layers.0.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00070-of-00131.safetensors"},"model.language_model.layers.25.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00071-of-00131.safetensors"},"model.language_model.layers.25.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.25.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.25.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.25.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.25.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.25.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.25.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.25.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.25.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.25.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00072-of-00131.safetensors"},"mtp.layers.0.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00072-of-00131.safetensors"},"mtp.layers.0.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00072-of-00131.safetensors"},"model.language_model.layers.26.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00073-of-00131.safetensors"},"model.language_model.layers.26.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.26.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.26.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.26.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.26.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.26.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.26.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.26.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.26.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.26.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.27.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.27.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.27.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.27.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00074-of-00131.safetensors"},"model.language_model.layers.27.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00075-of-00131.safetensors"},"model.language_model.layers.27.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.27.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00076-of-00131.safetensors"},"mtp.layers.0.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00076-of-00131.safetensors"},"model.language_model.layers.28.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00077-of-00131.safetensors"},"model.language_model.layers.28.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.28.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.28.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.28.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.28.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.28.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.28.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.28.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.28.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.28.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00078-of-00131.safetensors"},"mtp.layers.0.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00078-of-00131.safetensors"},"model.language_model.layers.29.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00079-of-00131.safetensors"},"model.language_model.layers.29.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.29.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.29.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.29.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.29.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.29.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.29.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.29.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.29.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.29.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.3.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.3.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.3.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.3.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00080-of-00131.safetensors"},"model.language_model.layers.3.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00081-of-00131.safetensors"},"model.language_model.layers.3.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.3.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00082-of-00131.safetensors"},"mtp.layers.0.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00082-of-00131.safetensors"},"model.language_model.layers.30.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00083-of-00131.safetensors"},"model.language_model.layers.30.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.30.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.30.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.30.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.30.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.30.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.30.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.30.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.30.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.30.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.31.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.31.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.31.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.31.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00084-of-00131.safetensors"},"model.language_model.layers.31.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00085-of-00131.safetensors"},"model.language_model.layers.31.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.31.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00086-of-00131.safetensors"},"mtp.layers.0.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00086-of-00131.safetensors"},"model.language_model.layers.32.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00087-of-00131.safetensors"},"model.language_model.layers.32.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.32.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.32.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.32.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.32.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.32.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.32.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.32.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.32.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.32.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00088-of-00131.safetensors"},"mtp.layers.0.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00088-of-00131.safetensors"},"model.language_model.layers.33.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00089-of-00131.safetensors"},"model.language_model.layers.33.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.33.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.33.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.33.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.33.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.33.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.33.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.33.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.33.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.33.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00090-of-00131.safetensors"},"mtp.layers.0.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00090-of-00131.safetensors"},"model.language_model.layers.34.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00091-of-00131.safetensors"},"model.language_model.layers.34.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.34.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.34.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.34.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.34.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.34.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.34.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.34.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.34.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.34.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.35.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.35.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.35.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.35.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00092-of-00131.safetensors"},"model.language_model.layers.35.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00093-of-00131.safetensors"},"model.language_model.layers.35.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.35.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00094-of-00131.safetensors"},"mtp.layers.0.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00094-of-00131.safetensors"},"model.language_model.layers.36.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00095-of-00131.safetensors"},"model.language_model.layers.36.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.36.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.36.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.36.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.36.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.36.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.36.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.36.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.36.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.36.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00096-of-00131.safetensors"},"model.language_model.layers.37.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00097-of-00131.safetensors"},"model.language_model.layers.37.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.37.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.37.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.37.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.37.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.37.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.37.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.37.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.37.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.37.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00098-of-00131.safetensors"},"model.language_model.layers.38.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00099-of-00131.safetensors"},"model.language_model.layers.38.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.38.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.38.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.38.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.38.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.38.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.38.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.38.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.38.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.38.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.39.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.39.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.39.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.39.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00100-of-00131.safetensors"},"mtp.layers.0.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00100-of-00131.safetensors"},"model.language_model.layers.39.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00101-of-00131.safetensors"},"model.language_model.layers.39.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.39.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00102-of-00131.safetensors"},"mtp.layers.0.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00102-of-00131.safetensors"},"model.language_model.layers.4.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00103-of-00131.safetensors"},"model.language_model.layers.4.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.4.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.4.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.4.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.4.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.4.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.4.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.4.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.4.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.4.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00104-of-00131.safetensors"},"model.language_model.layers.40.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00105-of-00131.safetensors"},"model.language_model.layers.40.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.40.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.40.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.40.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.40.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.40.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.40.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.40.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.40.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.40.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00106-of-00131.safetensors"},"mtp.layers.0.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00106-of-00131.safetensors"},"model.language_model.layers.41.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00107-of-00131.safetensors"},"model.language_model.layers.41.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.41.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.41.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.41.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.41.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.41.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.41.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.41.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.41.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.41.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00108-of-00131.safetensors"},"mtp.layers.0.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00108-of-00131.safetensors"},"mtp.layers.0.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00108-of-00131.safetensors"},"mtp.layers.0.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00108-of-00131.safetensors"},"model.language_model.layers.42.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00109-of-00131.safetensors"},"model.language_model.layers.42.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.42.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.42.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.42.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.42.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.42.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.42.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.42.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.42.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.42.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.43.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.43.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.43.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.43.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00110-of-00131.safetensors"},"model.language_model.layers.43.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00111-of-00131.safetensors"},"model.language_model.layers.43.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.43.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00112-of-00131.safetensors"},"mtp.layers.0.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00112-of-00131.safetensors"},"model.language_model.layers.44.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00113-of-00131.safetensors"},"model.language_model.layers.44.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.44.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.44.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.44.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.44.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.44.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.44.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.44.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.44.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.44.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00114-of-00131.safetensors"},"mtp.pre_fc_norm_embedding.weight":{"shape":[2560],"dtype":"BF16","numel":2560,"shard":"model-00114-of-00131.safetensors"},"model.language_model.layers.45.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00115-of-00131.safetensors"},"model.language_model.layers.45.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.45.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.45.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.45.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.45.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.45.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.45.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.45.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.45.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.45.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00116-of-00131.safetensors"},"model.language_model.layers.46.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00117-of-00131.safetensors"},"model.language_model.layers.46.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.46.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.46.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.46.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.46.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.46.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.46.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.46.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.46.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.46.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.47.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.47.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.47.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.47.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00118-of-00131.safetensors"},"mtp.pre_fc_norm_hidden.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00118-of-00131.safetensors"},"model.language_model.layers.47.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00119-of-00131.safetensors"},"model.language_model.layers.47.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.47.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00120-of-00131.safetensors"},"mtp.fc_embedding.weight":{"shape":[2560,2560],"dtype":"BF16","numel":6553600,"shard":"model-00120-of-00131.safetensors"},"model.language_model.layers.5.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00121-of-00131.safetensors"},"model.language_model.layers.5.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.5.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.5.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.5.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.5.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.5.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.5.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.5.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.5.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.5.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00122-of-00131.safetensors"},"model.language_model.layers.6.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00123-of-00131.safetensors"},"model.language_model.layers.6.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.6.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.6.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.6.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.6.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.6.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.6.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.6.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.6.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.6.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.7.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.7.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.7.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.7.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00124-of-00131.safetensors"},"mtp.fc_hidden.weight":{"shape":[2560,2560],"dtype":"BF16","numel":6553600,"shard":"model-00124-of-00131.safetensors"},"model.language_model.layers.7.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00125-of-00131.safetensors"},"model.language_model.layers.7.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.self_attn.indexer.index_qk_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.self_attn.indexer.k_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.self_attn.indexer.q_layernorm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.self_attn.k_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.self_attn.k_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.self_attn.o_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.self_attn.q_norm.weight":{"shape":[256],"dtype":"BF16","numel":256,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.self_attn.q_proj.weight":{"shape":[12288,2560],"dtype":"BF16","numel":31457280,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.7.self_attn.v_proj.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00126-of-00131.safetensors"},"model.language_model.layers.8.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00127-of-00131.safetensors"},"model.language_model.layers.8.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.8.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.8.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.8.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.8.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.8.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.8.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.8.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.8.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.8.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.attn_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.attn_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.attn_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.attn_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.linear_attn.A_log":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.linear_attn.conv1d.weight":{"shape":[10240,1,4],"dtype":"BF16","numel":40960,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.linear_attn.dt_bias":{"shape":[48],"dtype":"BF16","numel":48,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.linear_attn.in_proj_a.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.linear_attn.in_proj_b.weight":{"shape":[48,2560],"dtype":"BF16","numel":122880,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.linear_attn.in_proj_qkv.weight":{"shape":[10240,2560],"dtype":"BF16","numel":26214400,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.linear_attn.in_proj_z.weight":{"shape":[6144,2560],"dtype":"BF16","numel":15728640,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.linear_attn.norm.weight":{"shape":[128],"dtype":"BF16","numel":128,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.linear_attn.out_proj.weight":{"shape":[2560,6144],"dtype":"BF16","numel":15728640,"shard":"model-00128-of-00131.safetensors"},"model.language_model.layers.9.mlp.experts.gate_up_proj":{"shape":[512,1280,2560],"dtype":"BF16","numel":1677721600,"shard":"model-00129-of-00131.safetensors"},"model.language_model.embed_tokens.weight":{"shape":[248320,2560],"dtype":"BF16","numel":635699200,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp.experts.down_proj":{"shape":[512,2560,640],"dtype":"BF16","numel":838860800,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp.gate.weight":{"shape":[512,2560],"dtype":"BF16","numel":1310720,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp.shared_expert.down_proj.weight":{"shape":[2560,640],"dtype":"BF16","numel":1638400,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp.shared_expert.gate_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp.shared_expert.up_proj.weight":{"shape":[640,2560],"dtype":"BF16","numel":1638400,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp.shared_expert_gate.weight":{"shape":[1,2560],"dtype":"BF16","numel":2560,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp_hyper_connection.block_inject_weight.weight":{"shape":[4,10240],"dtype":"BF16","numel":40960,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp_hyper_connection.hc_norm.weight":{"shape":[10240],"dtype":"BF16","numel":10240,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp_hyper_connection.input_mix_weight_down.weight":{"shape":[320,10240],"dtype":"BF16","numel":3276800,"shard":"model-00130-of-00131.safetensors"},"model.language_model.layers.9.mlp_hyper_connection.input_mix_weight_up.weight":{"shape":[10240,320],"dtype":"BF16","numel":3276800,"shard":"model-00130-of-00131.safetensors"},"lm_head.weight":{"shape":[248320,2560],"dtype":"BF16","numel":635699200,"shard":"model-00131-of-00131.safetensors"}},"config":{"architectures":["Qwen4ExpForConditionalGeneration"],"image_token_id":248056,"language_model_only":false,"model_type":"qwen4_exp","text_config":{"attention_bias":false,"attention_dropout":0.0,"bos_token_id":248044,"dtype":"bfloat16","eos_token_id":248044,"full_attention_interval":4,"hc_count":4,"hc_lowrank":320,"head_dim":256,"heads_per_ngram":8,"hidden_act":"silu","hidden_size":2560,"indexer_budget":2048,"indexer_compress_ratio":4,"indexer_head_dim":128,"indexer_kv_heads":1,"indexer_n_heads":4,"initializer_range":0.02,"layer_types":["linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention","linear_attention","linear_attention","linear_attention","full_attention"],"linear_conv_kernel_dim":4,"linear_key_head_dim":128,"linear_num_key_heads":16,"linear_num_value_heads":48,"linear_value_head_dim":128,"make_ngram_vocab_size_divisible_by":128,"mamba_ssm_dtype":"float32","max_position_embeddings":262144,"model_type":"qwen4_exp_text","moe_intermediate_size":640,"mtp":{"hybrid":true,"layer_types":["full_attention"],"mtp_use_hidden_state_from_layer":null,"num_hidden_layers":1,"rope_theta":10000000},"mtp_num_hidden_layers":1,"mtp_use_dedicated_embeddings":false,"ngram_size":3,"ngram_vocab_size_base":20000000,"num_attention_heads":24,"num_experts":512,"num_experts_per_tok":10,"num_hidden_layers":48,"num_key_value_heads":2,"output_gate_type":"sigmoid","output_router_logits":false,"pad_token_id":null,"partial_rotary_factor":0.25,"ple_conv_kernel_size":4,"ple_embed_dim":2560,"ple_layer_ids":[2],"rms_norm_eps":1e-6,"rope_parameters":{"mrope_interleaved":true,"mrope_section":[11,11,10],"partial_rotary_factor":0.25,"rope_theta":10000000,"rope_type":"default"},"router_aux_loss_coef":0.001,"shared_expert_intermediate_size":640,"split_ngram_parts":128,"tie_word_embeddings":false,"use_cache":true,"vocab_size":248320},"tie_word_embeddings":false,"transformers_version":"5.8.0.dev0","video_token_id":248057,"vision_config":{"deepstack_visual_indexes":[],"depth":27,"hidden_act":"gelu_pytorch_tanh","hidden_size":1152,"in_channels":3,"initializer_range":0.02,"intermediate_size":4304,"model_type":"qwen4_exp","num_heads":16,"num_position_embeddings":2304,"out_hidden_size":2560,"patch_size":16,"spatial_merge_size":2,"temporal_patch_size":2},"vision_end_token_id":248054,"vision_start_token_id":248053}}