diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,2929 @@ +{ + "metadata": { + "ParamSize": 243, + "ParamBytes": 962793472.0, + "BitsPerParam": 4.500674339093837 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 50331648, + "records": [ + { + "name": "model.embed_tokens.q_weight", + "shape": [ + 49152, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 50331648, + "byteOffset": 0 + } + ], + "md5sum": "c2f2fec4ac63505a73e984e7be998001" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 32509952, + "records": [ + { + "name": "model.embed_tokens.q_scale", + "shape": [ + 49152, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 6291456 + }, + { + "name": "model.layers.0.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 6295552 + }, + { + "name": "model.layers.0.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 14684160 + }, + { + "name": "model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 15732736 + } + ], + "md5sum": "7fb0c857c7119ef2c42a35e51f45a049" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.1.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "81342b605812fe6f181e10a94d8aa2d1" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 32522240, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 0 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 2097152 + }, + { + "name": "model.layers.0.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 2101248 + }, + { + "name": "model.layers.0.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 8392704 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 9179136 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 11276288 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 11538432 + }, + { + "name": "model.layers.1.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 11542528 + }, + { + "name": "model.layers.1.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 19931136 + }, + { + "name": "model.layers.1.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 20979712 + }, + { + "name": "model.layers.1.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 23076864 + }, + { + "name": "model.layers.1.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 23080960 + }, + { + "name": "model.layers.1.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 29372416 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 30158848 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 32256000 + }, + { + "name": "model.layers.10.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 32518144 + } + ], + "md5sum": "20cc1062eb54c11d9aae9e7abf8ef7be" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.10.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.10.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.10.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.10.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.10.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "b3d08dbea8b63b6780aad833107aecc0" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "02311dfb84d5343314dd190a7bb57b25" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.10.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.10.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.11.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.11.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.11.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.11.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.11.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.11.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.11.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.12.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "0eea6f7a94d9811254b38922af5f4547" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.12.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.12.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.12.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.12.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.12.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "4cf8a21d894851fb7e3f92657e22200c" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "0047cd07d809031e8507dee7c3de54d6" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.12.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.12.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.13.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.13.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.13.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.13.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.13.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.13.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.13.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.14.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "02cfaf279d28fd906f8ee0239a7b65bc" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.14.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.14.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.14.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.14.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.14.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "76d4492ecbd35fb8c38c9435f70b8459" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "96c929de5728f39129cf8ea62e721080" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.14.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.14.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.15.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.15.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.15.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.15.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.15.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.15.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.15.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.16.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "09969f401cadbcfec597a004728ca292" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.16.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.16.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.16.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.16.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.16.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "feb309bbd5ee10c6bca2d6466a1ac2da" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "2fc5818d5fa409e34c7b5c5c6f5d1fba" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.16.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.16.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.17.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.17.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.17.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.17.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.17.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.17.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.17.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.18.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "030e18f221c36f0f748b1af629e5036e" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.18.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.18.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.18.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.18.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.18.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "f08edd37f1a69b445a5b7a15dfd08c27" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "c72b779e29878ca1bfb266b7f54d3f5f" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.18.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.18.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.19.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.19.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.19.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.19.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.19.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.19.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.19.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.2.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "6cb4be7ae7e58dd2d477289bc6f5a46f" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.2.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.2.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.2.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.2.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.2.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "c299fec07407ee987733a8459a7def05" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "a35b0559c6242215df96ec7e8a7e6298" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.2.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.2.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.20.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.20.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.20.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.20.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.20.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.20.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.20.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.21.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "d02eae8bb918ba91d5f4a17c77e0338b" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.21.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.21.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.21.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.21.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.21.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "4a08a3b63c6fd2359f1701a966751820" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "c10283e12d4fa81cff8cc1c1cd5f8dcf" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.21.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.21.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.22.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.22.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.22.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.22.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.22.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.22.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.22.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.23.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "2884113bbfcbde6815ad44db32a09586" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.23.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.23.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.23.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.23.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.23.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "8c875bf38049afb87d10434f0b665302" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "bff74b9fb5a174c7b671e43dad2ca901" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.23.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.23.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.3.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.3.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.3.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.3.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.3.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.3.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.3.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.4.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "751d4ba066f4f97d3e77072a09f1fb67" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.4.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.4.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.4.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.4.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.4.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "90dff5d06dbeeca23bd7e12059028908" + }, + { + "dataPath": "params_shard_29.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "7de5cfba9a6f9656591a650c03da7d26" + }, + { + "dataPath": "params_shard_30.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.4.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.4.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.5.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.5.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.5.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.5.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.5.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.5.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.5.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.6.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "df9523e386df12aa5c3dd177fae0eeb4" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.6.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.6.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.6.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.6.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.6.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "08b0395c4c0d180732208513950c343b" + }, + { + "dataPath": "params_shard_32.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "d82377ef68ae28a3fe09f4d43f820cfb" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.6.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.6.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.7.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.7.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.7.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.7.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.7.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.7.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.7.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.layers.8.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "0a7c9ea1f1b35d9c2053e47c5b2a7370" + }, + { + "dataPath": "params_shard_34.bin", + "format": "raw-shard", + "nbytes": 28315648, + "records": [ + { + "name": "model.layers.8.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 0 + }, + { + "name": "model.layers.8.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 8388608 + }, + { + "name": "model.layers.8.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 9437184 + }, + { + "name": "model.layers.8.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26214400 + }, + { + "name": "model.layers.8.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28311552 + } + ], + "md5sum": "f895186c0a432699b1fa567f59f9a71e" + }, + { + "dataPath": "params_shard_35.bin", + "format": "raw-shard", + "nbytes": 16777216, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 16777216, + "byteOffset": 0 + } + ], + "md5sum": "898d7732a1fb08b4812627c93a15c751" + }, + { + "dataPath": "params_shard_36.bin", + "format": "raw-shard", + "nbytes": 30420992, + "records": [ + { + "name": "model.layers.8.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "model.layers.8.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7077888 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9175040 + }, + { + "name": "model.layers.9.input_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 9437184 + }, + { + "name": "model.layers.9.mlp.down_proj.q_weight", + "shape": [ + 2048, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 8388608, + "byteOffset": 9441280 + }, + { + "name": "model.layers.9.mlp.down_proj.q_scale", + "shape": [ + 2048, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1048576, + "byteOffset": 17829888 + }, + { + "name": "model.layers.9.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18878464 + }, + { + "name": "model.layers.9.post_attention_layernorm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 20975616 + }, + { + "name": "model.layers.9.self_attn.qkv_proj.q_weight", + "shape": [ + 6144, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 20979712 + }, + { + "name": "model.layers.9.self_attn.qkv_proj.q_scale", + "shape": [ + 6144, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 27271168 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_weight", + "shape": [ + 2048, + 256 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28057600 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_scale", + "shape": [ + 2048, + 64 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30154752 + }, + { + "name": "model.norm.weight", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 30416896 + } + ], + "md5sum": "ac95253e6b621029e711cc228a4ce95d" + } + ] +} \ No newline at end of file