{ "metadata": { "ParamSize": 198, "ParamBytes": 3087428608.0, "BitsPerParam": 13.89882147873375 }, "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": "fac6ce8ba0d8ae438d10b5bb546c7cbd" }, { "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": "b0c672ae0f591da2d49f4dce650e72ae" }, { "dataPath": "params_shard_2.bin", "format": "raw-shard", "nbytes": 27525120, "records": [ { "name": "model.layers.0.mlp.down_proj.weight", "shape": [ 1536, 8960 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 27525120, "byteOffset": 0 } ], "md5sum": "db1e935ff9f5b6ff5da80f8b6d1628ff" }, { "dataPath": "params_shard_3.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": "baee6aadc8be0f23324126937a4cf15b" }, { "dataPath": "params_shard_4.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": "9581d907a6edac4b6bc1c357d0d2343f" }, { "dataPath": "params_shard_5.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": "a8b915076ac9a72930fb6871e62d46ef" }, { "dataPath": "params_shard_6.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": "972150741a85989c34825d9c970f6741" }, { "dataPath": "params_shard_7.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.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.0.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.0.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.0.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "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.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 17311744 }, { "name": "model.layers.1.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17315840 }, { "name": "model.layers.1.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22034432 }, { "name": "model.layers.1.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22037504 }, { "name": "model.layers.2.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.2.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 28332032 }, { "name": "model.layers.2.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28336128 }, { "name": "model.layers.2.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33054720 }, { "name": "model.layers.2.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33057792 } ], "md5sum": "cc217b23db4df89fcfed2ffa54ee328a" }, { "dataPath": "params_shard_8.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": "969f8fa80cdf5b272c2e607a2119c4e5" }, { "dataPath": "params_shard_9.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": "d860175d6e9380a30cfbe06262104ec7" }, { "dataPath": "params_shard_10.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": "3f5dad1ba616506c4e5084bb18bbf1d6" }, { "dataPath": "params_shard_11.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": "b993af8733b274d5e5aaacdc6462f74f" }, { "dataPath": "params_shard_12.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": "2877a35f91a04b1c58f6a824e8d61e87" }, { "dataPath": "params_shard_13.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": "8bad654a3c485e170a6f0e5ddab81db0" }, { "dataPath": "params_shard_14.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.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.3.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.3.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.3.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "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.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 17311744 }, { "name": "model.layers.4.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17315840 }, { "name": "model.layers.4.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22034432 }, { "name": "model.layers.4.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22037504 }, { "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.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 28332032 }, { "name": "model.layers.5.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28336128 }, { "name": "model.layers.5.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33054720 }, { "name": "model.layers.5.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33057792 } ], "md5sum": "9a67ac0bf8178e4054e781eb27f403ea" }, { "dataPath": "params_shard_15.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": "ccd5502649f7168e48df2ccbaeeb1d9b" }, { "dataPath": "params_shard_16.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": "271cf1aad384c2d628d1aaf62a140a9c" }, { "dataPath": "params_shard_17.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": "901c1912a0418111b928559f750e0d95" }, { "dataPath": "params_shard_18.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": "379946453685b3ee0b55935c9538eecd" }, { "dataPath": "params_shard_19.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": "56fa2871d8315eef920e5df9970ea9fe" }, { "dataPath": "params_shard_20.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": "817b9f5a7e095c5f05876994890301b4" }, { "dataPath": "params_shard_21.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.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.6.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.6.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.6.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "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.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 17311744 }, { "name": "model.layers.7.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17315840 }, { "name": "model.layers.7.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22034432 }, { "name": "model.layers.7.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22037504 }, { "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.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 28332032 }, { "name": "model.layers.8.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28336128 }, { "name": "model.layers.8.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33054720 }, { "name": "model.layers.8.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33057792 } ], "md5sum": "bcd5c8951311cb8debd10c7a9cbd2764" }, { "dataPath": "params_shard_22.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": "faf5cbdd3c2ff11883ce9f8554bcf498" }, { "dataPath": "params_shard_23.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": "17ae02bef8401386b5eb9f653b2f420a" }, { "dataPath": "params_shard_24.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": "69af3fefd0ad33870219cab220d98c1b" }, { "dataPath": "params_shard_25.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": "5862aaa3d0c9514ced28f0c391b70d8c" }, { "dataPath": "params_shard_26.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": "221ad0e352c9a06fbd95fa8990829aac" }, { "dataPath": "params_shard_27.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": "605ca260375f758e22d4265fffa356fc" }, { "dataPath": "params_shard_28.bin", "format": "raw-shard", "nbytes": 33060864, "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.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.9.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.9.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.9.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "name": "model.layers.10.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.10.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 17311744 }, { "name": "model.layers.10.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17315840 }, { "name": "model.layers.10.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22034432 }, { "name": "model.layers.10.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22037504 }, { "name": "model.layers.11.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.11.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 28332032 }, { "name": "model.layers.11.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28336128 }, { "name": "model.layers.11.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33054720 }, { "name": "model.layers.11.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33057792 } ], "md5sum": "03e434a9483fb2aaa19f4cab3e4fb083" }, { "dataPath": "params_shard_29.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": "dc39bbf30e260c4170b068cc8553d7cd" }, { "dataPath": "params_shard_30.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": "d6ebf5371ced23357eb54793db3a4afb" }, { "dataPath": "params_shard_31.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": "b23a6222ce16baad47b61fe41f8660b0" }, { "dataPath": "params_shard_32.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": "f579126381720332c477dfeb50701e20" }, { "dataPath": "params_shard_33.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": "44b78e073ac8337e052b9744380ec3c7" }, { "dataPath": "params_shard_34.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": "d205aa713ee99dee09b571cfedc7cefd" }, { "dataPath": "params_shard_35.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.12.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.12.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.12.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.12.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.12.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "name": "model.layers.13.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.13.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 17311744 }, { "name": "model.layers.13.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17315840 }, { "name": "model.layers.13.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22034432 }, { "name": "model.layers.13.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22037504 }, { "name": "model.layers.14.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.14.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 28332032 }, { "name": "model.layers.14.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28336128 }, { "name": "model.layers.14.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33054720 }, { "name": "model.layers.14.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33057792 } ], "md5sum": "54c052279cd7ffad4678b87826192f76" }, { "dataPath": "params_shard_36.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": "0dac5ce31c1f2eaa17a56d0da55f7b6d" }, { "dataPath": "params_shard_37.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": "6b9f7819701341646bc2a266108f95b2" }, { "dataPath": "params_shard_38.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": "04cd76c5b6fc4ebe6baaa8644b6930fd" }, { "dataPath": "params_shard_39.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": "349f0c70d75d8a410f1b236fd7e395b5" }, { "dataPath": "params_shard_40.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": "63960103006f54f63a378784156e1ff6" }, { "dataPath": "params_shard_41.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": "0f07069c62ca1e2d8c3e7ee712fe82b0" }, { "dataPath": "params_shard_42.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.15.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.15.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.15.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.15.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.15.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "name": "model.layers.16.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.16.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 17311744 }, { "name": "model.layers.16.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17315840 }, { "name": "model.layers.16.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22034432 }, { "name": "model.layers.16.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22037504 }, { "name": "model.layers.17.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.17.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 28332032 }, { "name": "model.layers.17.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28336128 }, { "name": "model.layers.17.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33054720 }, { "name": "model.layers.17.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33057792 } ], "md5sum": "cbdd25fd19637b648dff1256b0362d00" }, { "dataPath": "params_shard_43.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": "639c98457bec8c7e83c2c1da8b9410f2" }, { "dataPath": "params_shard_44.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": "c335e79b08a2fe03d63fa514b658d909" }, { "dataPath": "params_shard_45.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": "6e0525006bc32f3ec81492080dc9f7f2" }, { "dataPath": "params_shard_46.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": "d968c5af427bccf73fb824e0be561cd9" }, { "dataPath": "params_shard_47.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": "80e8b32cb11f59a6022acf852eea7ace" }, { "dataPath": "params_shard_48.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": "c075bcffc12a1d712ca76c1471ae444a" }, { "dataPath": "params_shard_49.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.18.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.18.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.18.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.18.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.18.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "name": "model.layers.19.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.19.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 17311744 }, { "name": "model.layers.19.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17315840 }, { "name": "model.layers.19.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22034432 }, { "name": "model.layers.19.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22037504 }, { "name": "model.layers.20.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.20.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 28332032 }, { "name": "model.layers.20.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28336128 }, { "name": "model.layers.20.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33054720 }, { "name": "model.layers.20.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33057792 } ], "md5sum": "73cd2787ae936b6aa7f0c3c9c85b3ff2" }, { "dataPath": "params_shard_50.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": "136a14789a41480fcc08d89e66ca2501" }, { "dataPath": "params_shard_51.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": "a3f54fbb03f00cc32caaa8b0a5cb8394" }, { "dataPath": "params_shard_52.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": "2ffd1ebbe30c22f30f61a3d92c032d1b" }, { "dataPath": "params_shard_53.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": "2a50d409322f1488c49d5287dd21c3b8" }, { "dataPath": "params_shard_54.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": "89a049c13e34f896861d5989e7af37ca" }, { "dataPath": "params_shard_55.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": "60e61d289310f677b122617d45a9c7e4" }, { "dataPath": "params_shard_56.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.21.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.21.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.21.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.21.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.21.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "name": "model.layers.22.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.22.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 17311744 }, { "name": "model.layers.22.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17315840 }, { "name": "model.layers.22.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22034432 }, { "name": "model.layers.22.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22037504 }, { "name": "model.layers.23.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.23.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 28332032 }, { "name": "model.layers.23.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28336128 }, { "name": "model.layers.23.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33054720 }, { "name": "model.layers.23.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33057792 } ], "md5sum": "2859dd48c6a504a6ff5fc777a5d0eb01" }, { "dataPath": "params_shard_57.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": "00c023336d6889bb9d24bcb3762ff819" }, { "dataPath": "params_shard_58.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": "725c3d334464020a85016e629f58eda8" }, { "dataPath": "params_shard_59.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": "09c4056fb69cfcce7f2c9829ba155e28" }, { "dataPath": "params_shard_60.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": "898fa6e97b95ddc00a3a685714dd2056" }, { "dataPath": "params_shard_61.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": "ae5ce4d265237aae0fd576bb6f528dca" }, { "dataPath": "params_shard_62.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": "58375129ef74f23538d522ef09315b51" }, { "dataPath": "params_shard_63.bin", "format": "raw-shard", "nbytes": 33060864, "records": [ { "name": "model.layers.24.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.24.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.24.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.24.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.24.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "name": "model.layers.25.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 11020288 }, { "name": "model.layers.25.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 17311744 }, { "name": "model.layers.25.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 17315840 }, { "name": "model.layers.25.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22034432 }, { "name": "model.layers.25.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 22037504 }, { "name": "model.layers.26.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 22040576 }, { "name": "model.layers.26.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 28332032 }, { "name": "model.layers.26.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 28336128 }, { "name": "model.layers.26.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33054720 }, { "name": "model.layers.26.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 33057792 } ], "md5sum": "50113496b65c91d08a408ca51382050c" }, { "dataPath": "params_shard_64.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": "77bba71a9b67ed4377d409949e9ad145" }, { "dataPath": "params_shard_65.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": "0d6ad7eb791649a022f869b1edeb5355" }, { "dataPath": "params_shard_66.bin", "format": "raw-shard", "nbytes": 11023360, "records": [ { "name": "model.layers.27.self_attn.c_attn.weight", "shape": [ 2048, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 6291456, "byteOffset": 0 }, { "name": "model.layers.27.self_attn.c_attn.bias", "shape": [ 2048 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4096, "byteOffset": 6291456 }, { "name": "model.layers.27.self_attn.o_proj.weight", "shape": [ 1536, 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 4718592, "byteOffset": 6295552 }, { "name": "model.layers.27.input_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11014144 }, { "name": "model.layers.27.post_attention_layernorm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11017216 }, { "name": "model.norm.weight", "shape": [ 1536 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3072, "byteOffset": 11020288 } ], "md5sum": "29c619ad6174a67fe28726e26cc3974d" } ] }