{ "metadata": { "ParamSize": 198, "ParamBytes": 3087428608.0, "BitsPerParam": 16.0 }, "records": [ { "dataPath": "params_shard_0.bin", "format": "raw-shard", "nbytes": 466747392, "records": [ { "name": "model.embed_tokens.weight", "shape": [ 151936, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 466747392, "byteOffset": 0 } ], "md5sum": "dc5478854be743693cef7917d1e7ce4c" }, { "dataPath": "params_shard_1.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.0.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "e067930d718e625798bc215501ef772e" }, { "dataPath": "params_shard_2.bin", "format": "raw-shard", "nbytes": 27535360, "records": [ { "name": "model.layers.0.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 0 }, { "name": "model.layers.0.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 3072 }, { "name": "model.layers.0.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 27528192 }, { "name": "model.layers.0.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 27531264 } ], "md5sum": "6e1ea98a9094cc0eb683d5027d453b18" }, { "dataPath": "params_shard_3.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.1.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "ddba98c52941cd30517222e7a1e744ec" }, { "dataPath": "params_shard_4.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.1.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "b43ee04337f01512b2699484f53889e2" }, { "dataPath": "params_shard_5.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.10.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "6f468176df511a1de802eb813dab6b80" }, { "dataPath": "params_shard_6.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.10.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "bf7baca2f207dbaa9b6a2e9bc06fae28" }, { "dataPath": "params_shard_7.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.11.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "b22c5b004d6360c776f1050766a39018" }, { "dataPath": "params_shard_8.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.11.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "2ab3cc8300727c2e1f7da4c5e08e9b0d" }, { "dataPath": "params_shard_9.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.0.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.0.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.layers.1.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 }, { "name": "model.layers.1.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11013120 }, { "name": "model.layers.1.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 11016192 }, { "name": "model.layers.1.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.1.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17311744 }, { "name": "model.layers.10.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22030336 }, { "name": "model.layers.10.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22033408 }, { "name": "model.layers.10.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 22036480 }, { "name": "model.layers.10.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.10.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28332032 }, { "name": "model.layers.11.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33050624 }, { "name": "model.layers.11.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33053696 }, { "name": "model.layers.11.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 33056768 } ], "md5sum": "3c2d0b091ca815935b4e767fc456152a" }, { "dataPath": "params_shard_10.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.12.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "8640e2a772e5f505cff9ef4d1278b993" }, { "dataPath": "params_shard_11.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.12.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "1a13ad3c671d922d98496eccd823b010" }, { "dataPath": "params_shard_12.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.13.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "04b10eb2f8d045366d4c14b82c37d12d" }, { "dataPath": "params_shard_13.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.13.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "6350144d0f2d5da6cc69725ace1f47c4" }, { "dataPath": "params_shard_14.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.14.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "78cc8750968e5324e5c549bfe82efa7f" }, { "dataPath": "params_shard_15.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.14.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "f8c5b627e8354e0cafa8f690c88b0d77" }, { "dataPath": "params_shard_16.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.11.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.11.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.layers.12.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 }, { "name": "model.layers.12.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11013120 }, { "name": "model.layers.12.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 11016192 }, { "name": "model.layers.12.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.12.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17311744 }, { "name": "model.layers.13.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22030336 }, { "name": "model.layers.13.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22033408 }, { "name": "model.layers.13.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 22036480 }, { "name": "model.layers.13.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.13.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28332032 }, { "name": "model.layers.14.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33050624 }, { "name": "model.layers.14.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33053696 }, { "name": "model.layers.14.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 33056768 } ], "md5sum": "df9950bdd911731e3dd85ddda77fd582" }, { "dataPath": "params_shard_17.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.15.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "7aa26bb0aeed376ea1898ad7cafaed13" }, { "dataPath": "params_shard_18.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.15.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "e47ac924fd3d52b09d7777158796354a" }, { "dataPath": "params_shard_19.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.16.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "5b7bc689b7dfec86ef5500b58aa9a350" }, { "dataPath": "params_shard_20.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.16.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "3ac080a9e359ed385700a8b5d6bfaff1" }, { "dataPath": "params_shard_21.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.17.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "4e2169393a9a12174945b636f734f32c" }, { "dataPath": "params_shard_22.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.17.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "8b6ce185e0fed51ec8c098d879feb573" }, { "dataPath": "params_shard_23.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.14.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.14.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.layers.15.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 }, { "name": "model.layers.15.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11013120 }, { "name": "model.layers.15.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 11016192 }, { "name": "model.layers.15.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.15.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17311744 }, { "name": "model.layers.16.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22030336 }, { "name": "model.layers.16.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22033408 }, { "name": "model.layers.16.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 22036480 }, { "name": "model.layers.16.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.16.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28332032 }, { "name": "model.layers.17.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33050624 }, { "name": "model.layers.17.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33053696 }, { "name": "model.layers.17.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 33056768 } ], "md5sum": "7c2462692938c376178ad1c00edab3ba" }, { "dataPath": "params_shard_24.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.18.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "e7ba6a56026eaab113b222ed28710af7" }, { "dataPath": "params_shard_25.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.18.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "58523fd8e8ea53ff4120e37738bbb08f" }, { "dataPath": "params_shard_26.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.19.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "d3bacbc04c8bb0d0a1c592196f2e5923" }, { "dataPath": "params_shard_27.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.19.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "906fda15b65c863eaa42e0d018287cd2" }, { "dataPath": "params_shard_28.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.2.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "5b444ecd6e9f163c2dc84e01e43e30b0" }, { "dataPath": "params_shard_29.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.2.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "5d14245af8728a641729575cb062050a" }, { "dataPath": "params_shard_30.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.17.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.17.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.layers.18.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 }, { "name": "model.layers.18.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11013120 }, { "name": "model.layers.18.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 11016192 }, { "name": "model.layers.18.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.18.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17311744 }, { "name": "model.layers.19.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22030336 }, { "name": "model.layers.19.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22033408 }, { "name": "model.layers.19.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 22036480 }, { "name": "model.layers.19.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.19.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28332032 }, { "name": "model.layers.2.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33050624 }, { "name": "model.layers.2.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33053696 }, { "name": "model.layers.2.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 33056768 } ], "md5sum": "047565e1ead6ed8171acb0bb05dfe4f6" }, { "dataPath": "params_shard_31.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.20.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "e1920edf5a425577488f56494f8c6799" }, { "dataPath": "params_shard_32.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.20.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "61841b982921f56767fd49c20571f051" }, { "dataPath": "params_shard_33.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.21.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "93a633e61cada747ea483b5ada3e509c" }, { "dataPath": "params_shard_34.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.21.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "b69fb9fd8484d9f0574c48361a965cc5" }, { "dataPath": "params_shard_35.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.22.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "e2b90bd426da7a86cb80224a597ace2b" }, { "dataPath": "params_shard_36.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.22.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "fcaefdca80fe0773e56358c090a6ceb9" }, { "dataPath": "params_shard_37.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.2.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.2.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.layers.20.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 }, { "name": "model.layers.20.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11013120 }, { "name": "model.layers.20.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 11016192 }, { "name": "model.layers.20.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.20.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17311744 }, { "name": "model.layers.21.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22030336 }, { "name": "model.layers.21.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22033408 }, { "name": "model.layers.21.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 22036480 }, { "name": "model.layers.21.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.21.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28332032 }, { "name": "model.layers.22.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33050624 }, { "name": "model.layers.22.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33053696 }, { "name": "model.layers.22.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 33056768 } ], "md5sum": "2a96d0a5c817151eec586fceab0ab0b8" }, { "dataPath": "params_shard_38.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.23.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "f52e06df9574515f4877cb789944514c" }, { "dataPath": "params_shard_39.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.23.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "39f9ee9b27a38ea757802010872449ed" }, { "dataPath": "params_shard_40.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.24.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "6c7cce5cffb798ed35ae4d08df743fe0" }, { "dataPath": "params_shard_41.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.24.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "c7c785f203286280e3068eda1226b619" }, { "dataPath": "params_shard_42.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.25.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "9aa68fccf3110a34dd87bd6fa8cff3ef" }, { "dataPath": "params_shard_43.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.25.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "b775b1981eb39c67f4501dee796d4d82" }, { "dataPath": "params_shard_44.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.22.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.22.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.layers.23.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 }, { "name": "model.layers.23.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11013120 }, { "name": "model.layers.23.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 11016192 }, { "name": "model.layers.23.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.23.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17311744 }, { "name": "model.layers.24.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22030336 }, { "name": "model.layers.24.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22033408 }, { "name": "model.layers.24.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 22036480 }, { "name": "model.layers.24.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.24.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28332032 }, { "name": "model.layers.25.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33050624 }, { "name": "model.layers.25.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33053696 }, { "name": "model.layers.25.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 33056768 } ], "md5sum": "2a42cdaadb418cc8b34daf6a1b743ca8" }, { "dataPath": "params_shard_45.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.26.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "dde0bedc0f713bf40e8ef1006a59b027" }, { "dataPath": "params_shard_46.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.26.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "b1a5630fceb2992d970f390df8e593e5" }, { "dataPath": "params_shard_47.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.27.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "ca2b36e5409ba1d0e8fc763531eec46c" }, { "dataPath": "params_shard_48.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.27.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "4d2db27561b873e6f523a1ac1fc431a7" }, { "dataPath": "params_shard_49.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.3.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "0fb7691571e7c645aa3a5f764664b64e" }, { "dataPath": "params_shard_50.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.3.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "240e2893ac2a56bf4d20d112d0cdfa03" }, { "dataPath": "params_shard_51.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.25.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.25.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.layers.26.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 }, { "name": "model.layers.26.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11013120 }, { "name": "model.layers.26.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 11016192 }, { "name": "model.layers.26.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.26.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17311744 }, { "name": "model.layers.27.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22030336 }, { "name": "model.layers.27.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22033408 }, { "name": "model.layers.27.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 22036480 }, { "name": "model.layers.27.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.27.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28332032 }, { "name": "model.layers.3.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33050624 }, { "name": "model.layers.3.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33053696 }, { "name": "model.layers.3.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 33056768 } ], "md5sum": "084acc443ccb211387199e34607196a0" }, { "dataPath": "params_shard_52.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.4.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "2599506f5f40fc8e24cd03cdab25d9ec" }, { "dataPath": "params_shard_53.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.4.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "377ddf1e6ef01df76df6f45b47dd7159" }, { "dataPath": "params_shard_54.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.5.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "00f283813acc0f97f57fc4a8b3186310" }, { "dataPath": "params_shard_55.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.5.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "baa28a19fd8d159da7395b7428563ff9" }, { "dataPath": "params_shard_56.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.6.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "156da00f0de6936cf3c72aa02ec72ec3" }, { "dataPath": "params_shard_57.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.6.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "c42214ad3ca08e10103412dc630f3e9c" }, { "dataPath": "params_shard_58.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.3.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.3.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.layers.4.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 }, { "name": "model.layers.4.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11013120 }, { "name": "model.layers.4.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 11016192 }, { "name": "model.layers.4.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.4.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17311744 }, { "name": "model.layers.5.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22030336 }, { "name": "model.layers.5.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22033408 }, { "name": "model.layers.5.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 22036480 }, { "name": "model.layers.5.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.5.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28332032 }, { "name": "model.layers.6.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33050624 }, { "name": "model.layers.6.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33053696 }, { "name": "model.layers.6.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 33056768 } ], "md5sum": "e3299b1d44b3430d346c3ba55e44ebac" }, { "dataPath": "params_shard_59.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.7.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "3e0814820f37bbe1657146617d343337" }, { "dataPath": "params_shard_60.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.7.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "aacdd7799f2ddbf41c21dbebe79798b9" }, { "dataPath": "params_shard_61.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.8.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "6c11341814ae3390103fd06a0eb3c90e" }, { "dataPath": "params_shard_62.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.8.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "22f72910a0321e3237d1dd708141b6cc" }, { "dataPath": "params_shard_63.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.9.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "62088d4a8bc1900127923d23a29662a1" }, { "dataPath": "params_shard_64.bin", "format": "raw-shard", "nbytes": 55050240, "records": [ { "name": "model.layers.9.mlp.gate_up_proj.weight", "shape": [ 17920, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 55050240, "byteOffset": 0 } ], "md5sum": "5d7dfcbbd08112df17ac43c388486bf0" }, { "dataPath": "params_shard_65.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.6.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.6.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.layers.7.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 }, { "name": "model.layers.7.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11013120 }, { "name": "model.layers.7.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 11016192 }, { "name": "model.layers.7.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.7.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17311744 }, { "name": "model.layers.8.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22030336 }, { "name": "model.layers.8.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22033408 }, { "name": "model.layers.8.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 22036480 }, { "name": "model.layers.8.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.8.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28332032 }, { "name": "model.layers.9.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33050624 }, { "name": "model.layers.9.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33053696 }, { "name": "model.layers.9.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 33056768 } ], "md5sum": "ad8285600a0794e1830b5778b7a30050" }, { "dataPath": "params_shard_66.bin", "format": "raw-shard", "nbytes": 11013120, "records": [ { "name": "model.layers.9.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.9.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6291456 }, { "name": "model.norm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11010048 } ], "md5sum": "2cc3dc4cf44fd4ed07ad932d30783139" } ] }