| { | |
| "metadata": { | |
| "ParamSize": 226, | |
| "ParamBytes": 1192099840.0, | |
| "BitsPerParam": 16.0 | |
| }, | |
| "records": [ | |
| { | |
| "dataPath": "params_shard_0.bin", | |
| "format": "raw-shard", | |
| "nbytes": 311164928, | |
| "records": [ | |
| { | |
| "name": "model.embed_tokens.weight", | |
| "shape": [ | |
| 151936, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 311164928, | |
| "byteOffset": 0 | |
| } | |
| ], | |
| "md5sum": "fa578c5e93879960baed59a72b9e364c" | |
| }, | |
| { | |
| "dataPath": "params_shard_1.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31463936, | |
| "records": [ | |
| { | |
| "name": "model.layers.0.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.0.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 2048 | |
| }, | |
| { | |
| "name": "model.layers.0.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6293504 | |
| }, | |
| { | |
| "name": "model.layers.0.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.0.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18878464 | |
| }, | |
| { | |
| "name": "model.layers.0.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18878720 | |
| }, | |
| { | |
| "name": "model.layers.0.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27267328 | |
| }, | |
| { | |
| "name": "model.layers.0.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31461632 | |
| }, | |
| { | |
| "name": "model.layers.1.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31461888 | |
| } | |
| ], | |
| "md5sum": "59fbb7cdfff6def738f18e5805e43da5" | |
| }, | |
| { | |
| "dataPath": "params_shard_2.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.1.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.1.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.1.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.1.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.1.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.1.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.1.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.10.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "d0b695b7dbd1e3a08d098587f79c05c7" | |
| }, | |
| { | |
| "dataPath": "params_shard_3.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.10.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.10.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.10.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.10.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.10.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.10.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.10.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.11.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "23c6119cfe29690e03ddbca5a8bfa9e9" | |
| }, | |
| { | |
| "dataPath": "params_shard_4.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.11.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.11.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.11.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.11.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.11.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.11.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.11.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.12.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "64c33e0bbcbd459cdeaed8ee934eaebe" | |
| }, | |
| { | |
| "dataPath": "params_shard_5.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.12.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.12.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.12.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.12.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.12.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.12.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.12.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.13.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "d82bad38015ffff954383d8fe8b69d41" | |
| }, | |
| { | |
| "dataPath": "params_shard_6.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.13.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.13.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.13.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.13.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.13.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.13.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.13.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.14.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "bd6a393160580ac6dcd6f654abb78d34" | |
| }, | |
| { | |
| "dataPath": "params_shard_7.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.14.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.14.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.14.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.14.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.14.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.14.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.14.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.15.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "12c05ad5ee0b686f03dfe99f1a633d1b" | |
| }, | |
| { | |
| "dataPath": "params_shard_8.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.15.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.15.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.15.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.15.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.15.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.15.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.15.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.16.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "72408a4c200788c88d590dadda7f6472" | |
| }, | |
| { | |
| "dataPath": "params_shard_9.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.16.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.16.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.16.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.16.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.16.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.16.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.16.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.17.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "c950492468409963735628f39c36e0d2" | |
| }, | |
| { | |
| "dataPath": "params_shard_10.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.17.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.17.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.17.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.17.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.17.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.17.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.17.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.18.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "94b8389ef693730f15d568994b4c3dd0" | |
| }, | |
| { | |
| "dataPath": "params_shard_11.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.18.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.18.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.18.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.18.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.18.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.18.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.18.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.19.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "4991d758cda1af770758d9854eed4d9c" | |
| }, | |
| { | |
| "dataPath": "params_shard_12.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.19.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.19.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.19.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.19.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.19.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.19.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.19.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.2.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "89e3cdaea5ef7c0297080f8b116c9c63" | |
| }, | |
| { | |
| "dataPath": "params_shard_13.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.2.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.2.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.2.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.2.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.2.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.2.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.2.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.20.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "daa70c4c74da8cd6bad70acfed859a47" | |
| }, | |
| { | |
| "dataPath": "params_shard_14.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.20.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.20.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.20.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.20.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.20.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.20.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.20.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.21.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "1f3ca2e614aa566a9429768e245f7b6a" | |
| }, | |
| { | |
| "dataPath": "params_shard_15.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.21.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.21.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.21.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.21.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.21.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.21.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.21.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.22.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "9cc0d2b4307aebf1a02d54999c5258b4" | |
| }, | |
| { | |
| "dataPath": "params_shard_16.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.22.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.22.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.22.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.22.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.22.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.22.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.22.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.23.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "c6ee7305b6c81324b75f23d9056073ef" | |
| }, | |
| { | |
| "dataPath": "params_shard_17.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.23.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.23.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.23.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.23.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.23.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.23.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.23.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.24.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "abd65843e4746acd023cd09b9f7b1abb" | |
| }, | |
| { | |
| "dataPath": "params_shard_18.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.24.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.24.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.24.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.24.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.24.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.24.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.24.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.25.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "d2bc311fd38ae4ae7c84c3fa2c22be72" | |
| }, | |
| { | |
| "dataPath": "params_shard_19.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.25.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.25.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.25.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.25.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.25.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.25.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.25.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.26.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "5f1dfeee995b1523057b4c7766dd9d77" | |
| }, | |
| { | |
| "dataPath": "params_shard_20.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.26.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.26.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.26.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.26.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.26.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.26.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.26.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.27.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "96a8b7a941519f37e134515bff30e767" | |
| }, | |
| { | |
| "dataPath": "params_shard_21.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.27.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.27.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.27.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.27.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.27.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.27.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.27.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.3.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "f0baacced3f2df1c93723ee639d34c1b" | |
| }, | |
| { | |
| "dataPath": "params_shard_22.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.3.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.3.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.3.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.3.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.3.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.3.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.3.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.4.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "903f585acbff63ed14375ecb4ea8c37c" | |
| }, | |
| { | |
| "dataPath": "params_shard_23.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.4.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.4.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.4.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.4.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.4.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.4.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.4.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.5.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "d0c81570c5dfed167b40dae9011fe042" | |
| }, | |
| { | |
| "dataPath": "params_shard_24.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.5.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.5.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.5.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.5.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.5.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.5.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.5.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.6.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "bf4c9657dbf786512d382c5ea8f005a6" | |
| }, | |
| { | |
| "dataPath": "params_shard_25.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.6.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.6.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.6.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.6.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.6.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.6.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.6.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.7.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "e5fef9d76ede52c4260c0ebe39468872" | |
| }, | |
| { | |
| "dataPath": "params_shard_26.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.7.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.7.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.7.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.7.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.7.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.7.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.7.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.8.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "444418d4e10ca5b746cc82931542eb40" | |
| }, | |
| { | |
| "dataPath": "params_shard_27.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.8.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.8.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.8.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.8.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.8.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.8.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.8.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.layers.9.input_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "73553837f68aac1433b22bd1d6613a59" | |
| }, | |
| { | |
| "dataPath": "params_shard_28.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31461888, | |
| "records": [ | |
| { | |
| "name": "model.layers.9.mlp.down_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6291456, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.9.mlp.gate_up_proj.weight", | |
| "shape": [ | |
| 6144, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 12582912, | |
| "byteOffset": 6291456 | |
| }, | |
| { | |
| "name": "model.layers.9.post_attention_layernorm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 18874368 | |
| }, | |
| { | |
| "name": "model.layers.9.self_attn.k_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 18876416 | |
| }, | |
| { | |
| "name": "model.layers.9.self_attn.c_attn.weight", | |
| "shape": [ | |
| 4096, | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 8388608, | |
| "byteOffset": 18876672 | |
| }, | |
| { | |
| "name": "model.layers.9.self_attn.o_proj.weight", | |
| "shape": [ | |
| 1024, | |
| 2048 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4194304, | |
| "byteOffset": 27265280 | |
| }, | |
| { | |
| "name": "model.layers.9.self_attn.q_norm.weight", | |
| "shape": [ | |
| 128 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 256, | |
| "byteOffset": 31459584 | |
| }, | |
| { | |
| "name": "model.norm.weight", | |
| "shape": [ | |
| 1024 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 2048, | |
| "byteOffset": 31459840 | |
| } | |
| ], | |
| "md5sum": "60e4f93a66fc1a62d2384bd159cd68da" | |
| } | |
| ] | |
| } |