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": "6e0db1a93d993cffb7aaf8a9c1d4ce7c" + }, + { + "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": "bceb47f2e8a374913b29fab331a13f8c" + }, + { + "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": "8511be7877cd3c6e99f0f49426aefa6f" + }, + { + "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": "f80ca218c680b378488bc5e48967eaed" + }, + { + "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": "14e75bdeb3ca9740aec9082ef10b80e3" + }, + { + "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": "93d06a5a9e53c9a0a5e15a0d0aededa8" + }, + { + "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": "4801e2c9e2d1ac1b151b8438a1951600" + }, + { + "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": "ab011827b74165c31cb7e913000a03a5" + }, + { + "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": "d2a3af264bb95a606d322ccb71015fd1" + }, + { + "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": "b9284a51ed8efe48a123e0fab5cbdc4b" + }, + { + "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": "4df85b281e06fb643d0acd37ae8b6a71" + }, + { + "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": "40d8f88af23507abae9873f683240567" + }, + { + "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": "d1a14ed6cb27b036d95b2c888a1f3bfd" + }, + { + "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": "c3b3c75014af5ea5243c8ff74373becc" + }, + { + "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": "54b04ed880deaac353d30b4636d197d4" + }, + { + "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": "6d551e55d864d5f598202f8f3e23ade4" + }, + { + "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": "c0cc833b62dc1b302aa42953b0ad789e" + }, + { + "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": "f3ea948d828eadedff606ddedbfeb944" + }, + { + "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": "525697ee4bee4b774eb81361b7afdd2d" + }, + { + "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": "658fd588e7355f91691d0af2c131de9b" + }, + { + "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": "9e6b6de7a446784043b50a5488662fcf" + }, + { + "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": "fdfc916a56459bd9f743e0a4a331dd08" + }, + { + "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": "f5e20841a36cdfca9d78edec42b0a374" + }, + { + "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": "03cbd6be438f519a4ca74de8757f3534" + }, + { + "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": "7ef5a9a8a9bbff3bd1fb7dd079670251" + }, + { + "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": "ef4b07d5dbb0cd18f214a479670af05a" + }, + { + "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": "dbb6cf19479b0158c6da381b12acecfa" + }, + { + "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": "4527c6adb24660f12d6d19de44dc3274" + }, + { + "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": "cbaf83f606a41629188af082cef13f2e" + }, + { + "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": "e50c07948357246c0bcb75a61aabd420" + }, + { + "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": "6854eab3b4f184bb3e52eb7495cd50dd" + }, + { + "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": "90e99aa085c315d02b5d824f5ed89a2f" + }, + { + "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": "d564cf20ac6b76b916055aa303523d67" + }, + { + "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": "6b12fcdfadec536a3ba38225faa1cadd" + }, + { + "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": "63c5a8fe7db5e5a73f6b6a606a24d105" + }, + { + "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": "d0b9a8adf09b7225ebfe3654c1e36332" + }, + { + "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": "89f99eedb35b470ff72f218669a3da2c" + } + ] +} \ No newline at end of file