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": "7ffd8d620a5c437e5ee792a0e6e3e32b" + }, + { + "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": "1a90e7b078eeceda3103868820eb434d" + }, + { + "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": "2c5a5c7076eb882b0573b14c9eaf5fa5" + }, + { + "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": "c3d8e0b470d9b86af1c31591222ffc5f" + }, + { + "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": "3b336316e00786466b201ae0a42ae2c9" + }, + { + "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": "22052ea8a22f23687ab374fe79ace1c4" + }, + { + "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": "bf7f1ab8271f531ab995747368028f03" + }, + { + "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": "0f5860d2d3e3e789798da7312207cd1f" + }, + { + "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": "7f066345f1861a0b032e4479b4c27448" + }, + { + "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": "815ae34afc89bf2326b00f976b8e4d88" + }, + { + "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": "4bb37a0e22422db531162467ddbcd183" + }, + { + "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": "e1defca4aa622538801f762a056b257e" + }, + { + "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": "c42f42ee46ce14a6356223bf7cf8fceb" + }, + { + "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": "54b4b80a29b165d0cee279c499aa972d" + }, + { + "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": "d413a9946225c50923bb736594a3925d" + }, + { + "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": "bbb79606e76cc78df4ead21879de7a5d" + }, + { + "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": "8f4e65a524c3fc2ff75f1212f4f60e34" + }, + { + "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": "051f667dca321c46fd152b0178f44b85" + }, + { + "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": "31e211fe39cc18fc255ca4efe5963268" + }, + { + "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": "f939f21e8a4539be757e938a4e53b41d" + }, + { + "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": "089da0cdc4587afbee4b322008b8ba76" + }, + { + "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": "2f9e53f3139ca81b4ac7a00764fc3db8" + }, + { + "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": "b2af01d52f699898406b1a5c9e6a7e24" + }, + { + "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": "7ca6ef8dc21c7080e62f483d6ffd681b" + }, + { + "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": "f3bd7453b603462dc38ab320b359d6ae" + }, + { + "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": "8edec011a5ee8189393fa7f6670af74e" + }, + { + "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": "65a697e177b5b9f97ea1c879e3f582bc" + }, + { + "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": "f06f1ed2c38ae1e343c31dde53848f7b" + }, + { + "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": "b812fcf34cb1be5ba597f137af8e42aa" + }, + { + "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": "f76176e6b123164f356d7df079070bd0" + }, + { + "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": "e783197178dfe6a59ca622e8a8eb1942" + }, + { + "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": "5950e479b973625ba5ee93427cfd51ef" + }, + { + "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": "c3aa6d58e03afe3f7a923289ff15cde9" + }, + { + "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": "560b98fef74aed7b569595baf471becf" + }, + { + "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": "9998d3bcc4f88626d72d6dff7db60972" + }, + { + "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": "b21909fb410753bc7cfbbe62a44d3636" + }, + { + "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": "aa7445ee7fc24553056d77687632a266" + } + ] +} \ No newline at end of file