diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,10103 @@ +{ + "metadata": { + "ParamSize": 709, + "ParamBytes": 18431289344.0, + "BitsPerParam": 3.2800260261877354 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 389283840, + "records": [ + { + "name": "lm_head.q_weight", + "shape": [ + 152064, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 389283840, + "byteOffset": 0 + } + ], + "md5sum": "1812520e66bbf73e486385e783727dd3" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 48660480, + "records": [ + { + "name": "lm_head.q_scale", + "shape": [ + 152064, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 48660480, + "byteOffset": 0 + } + ], + "md5sum": "56c0849d089c4a4f54f0633bde71c27b" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.63.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "98d6de35e30b2d15222ff9a22b004000" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.63.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "b186ca3f732341e22a72da22107949b6" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 389283840, + "records": [ + { + "name": "model.embed_tokens.q_weight", + "shape": [ + 152064, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 389283840, + "byteOffset": 0 + } + ], + "md5sum": "98c00e3c4b026241e96cc4a0e656fb2c" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 48660480, + "records": [ + { + "name": "model.embed_tokens.q_scale", + "shape": [ + 152064, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 48660480, + "byteOffset": 0 + } + ], + "md5sum": "4d4b2e1b6b19c5450970e25df1e70b6f" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.0.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "48533e192c51c9ab3427f9ed5dd73cc3" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 26583040, + "records": [ + { + "name": "model.layers.63.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 0 + }, + { + "name": "model.layers.63.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 10240 + }, + { + "name": "model.layers.63.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8857600 + }, + { + "name": "model.layers.63.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26552320 + }, + { + "name": "model.norm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26562560 + }, + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26572800 + } + ], + "md5sum": "96fa740a71c70245aa92bcd188657ac5" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "7f2f6f78dda678f0f4da016e9e8502f4" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.0.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "092a887952d1cb0d5ce51dea5ac53d27" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.0.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.0.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.0.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "ae525753226650a65b38ed89ecd2d402" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.1.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "125fb7b3a69e98e01f7a45c230229306" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.1.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "9685a8bb4a9a9a858a7c73c945904ffd" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.1.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "e11f04b841220b7cc1e75a12246d2c3a" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.1.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "8ff6401047828e3159046dd8c1c00383" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.1.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.1.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.1.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.1.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "d6667a64240014076bfc12b1a21e7596" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.2.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "45217688c86773b5df31a5e815a74fa5" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "7b47cacdd85092242fc52fe19e93ed0a" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "79bec92719dde84aeb3808aeb917c84b" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.2.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "26a7db5897ccbd6eee8ccb6721929900" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.1.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.2.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.2.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.2.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.2.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.2.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "05e3c1be0c537dc677f029279d3745aa" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "8baca1331183a776233f7ae6847a38a7" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.3.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "73d149798827ba26bb310cac5c45ed08" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.2.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.3.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.3.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "37008143ff6efec899b20db94ef2521b" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.10.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "d7eb9c51822393cbe3904f2d61c5f23c" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.10.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "4f3be47222c13721961da42dd8376e68" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.10.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "dfc9748063862db89df9e90a8bdb33ca" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.10.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "cd1da33acf6589c8e2b92477524f0e56" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 28215296, + "records": [ + { + "name": "model.layers.3.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.10.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.10.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.10.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.10.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 25907200 + }, + { + "name": "model.layers.10.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 25921536 + } + ], + "md5sum": "0931f373358cc748d901008400604550" + }, + { + "dataPath": "params_shard_29.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.11.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "621684dc87a1618e91e3feaf9e96455e" + }, + { + "dataPath": "params_shard_30.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "968c9d4c53dec36d03172a21fb95257d" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "ff5c54d334b178153a67db9003761359" + }, + { + "dataPath": "params_shard_32.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.11.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "f812b61e3acc509527f810c2cf2c4ae4" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.10.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.11.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.11.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.11.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.11.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.11.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "b5100619952ed6ca5e20d90b863e6771" + }, + { + "dataPath": "params_shard_34.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.12.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "dcb08ba5d6ea0649b855bf0fb9a0a2e5" + }, + { + "dataPath": "params_shard_35.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.12.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "e5106a91c06ec2617ae36cdd52c00a5f" + }, + { + "dataPath": "params_shard_36.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.12.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "939d08a8ce92839a55f354b8a6d1a291" + }, + { + "dataPath": "params_shard_37.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.12.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "b6c098b888d06370d75060cb421ed145" + }, + { + "dataPath": "params_shard_38.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.11.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.12.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.12.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.12.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.12.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.12.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "cd65f85e7d0be2c7f5bdbdfb3d235732" + }, + { + "dataPath": "params_shard_39.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "33a7857cfce2d84999ac75904e1befe2" + }, + { + "dataPath": "params_shard_40.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.13.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "1b5fd495b2f5acb6aaaed50c25065659" + }, + { + "dataPath": "params_shard_41.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.12.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.13.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.13.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "486eab51a9953ab62603ab81eaf8ded4" + }, + { + "dataPath": "params_shard_42.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.8.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "ac68272301c5214c5e32e8452d7c02c4" + }, + { + "dataPath": "params_shard_43.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "35b2a66466674a565b48362154e5cb2a" + }, + { + "dataPath": "params_shard_44.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "d11bf88df718cae6ffe045d56b1bcc57" + }, + { + "dataPath": "params_shard_45.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.9.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "522ae62b513306d33618184b357720bb" + }, + { + "dataPath": "params_shard_46.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.13.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.8.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.8.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.8.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.9.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "eab9c8c8a26e3cac66e697dcaafbfc68" + }, + { + "dataPath": "params_shard_47.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "1fbbd83d80c65faae2bb6197e477a012" + }, + { + "dataPath": "params_shard_48.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.9.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "93323837f7795b95c278b850173b3379" + }, + { + "dataPath": "params_shard_49.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.9.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.9.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.9.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.9.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.9.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "f1ba4591b82f4f20e17cc18875f82126" + }, + { + "dataPath": "params_shard_50.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.13.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "92f8b4d8ac7db2d1a163a7c21e0b8fbe" + }, + { + "dataPath": "params_shard_51.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.14.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "40b1030a8708c349b383bbff98a11bc8" + }, + { + "dataPath": "params_shard_52.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.14.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "febe805284ad11cbbd378d90e976e6be" + }, + { + "dataPath": "params_shard_53.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.14.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "9e06c59f802783b49ca2859826569776" + }, + { + "dataPath": "params_shard_54.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.14.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "108c81f3749199e2d1542b2a9bdde0ad" + }, + { + "dataPath": "params_shard_55.bin", + "format": "raw-shard", + "nbytes": 32495616, + "records": [ + { + "name": "model.layers.9.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.13.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.13.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.13.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.14.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23613440 + }, + { + "name": "model.layers.14.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 23623680 + }, + { + "name": "model.layers.14.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 32471040 + }, + { + "name": "model.layers.14.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32481280 + } + ], + "md5sum": "39951ea478c4736c71f0969d3f9cf36d" + }, + { + "dataPath": "params_shard_56.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.15.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "44b6b0f49cf5dcabb45be86e19a42ca6" + }, + { + "dataPath": "params_shard_57.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "c043868e8ee1445c91023670f3bcc0ab" + }, + { + "dataPath": "params_shard_58.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "f93a06b305e12824704c146de6859df7" + }, + { + "dataPath": "params_shard_59.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.15.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "8f16000ad72f1e18a7a35dd1cae7dec2" + }, + { + "dataPath": "params_shard_60.bin", + "format": "raw-shard", + "nbytes": 28215296, + "records": [ + { + "name": "model.layers.14.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.15.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.15.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.15.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.15.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 25907200 + }, + { + "name": "model.layers.15.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 25921536 + } + ], + "md5sum": "1071fe52648dee2b5a00b0fdf149155e" + }, + { + "dataPath": "params_shard_61.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.16.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "f860c7718385bb05e4d0c949ab06b5c4" + }, + { + "dataPath": "params_shard_62.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.16.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "92e8cda591f97b212defc451daf81e02" + }, + { + "dataPath": "params_shard_63.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.16.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "41de997392f1358e3e2dcf5e99dd920a" + }, + { + "dataPath": "params_shard_64.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.16.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "500bf9081ae1b8cad2b363e4ae88dabf" + }, + { + "dataPath": "params_shard_65.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.15.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.16.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.16.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.16.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.16.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.16.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "5da73b29e253f4804553d4b16217b970" + }, + { + "dataPath": "params_shard_66.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.17.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "92f65085689bf52e700f0a7a0a64d2a7" + }, + { + "dataPath": "params_shard_67.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "2e484d5e1c17a38b30adfe5b07a391f6" + }, + { + "dataPath": "params_shard_68.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "5df4aa1ca861bb11d969be2960655f55" + }, + { + "dataPath": "params_shard_69.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.17.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "ca1c813b3333d2aa25b34094f5187368" + }, + { + "dataPath": "params_shard_70.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.16.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.17.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.17.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.17.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.17.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.17.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "d771c17c82d92739185d76e87a0c4beb" + }, + { + "dataPath": "params_shard_71.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.18.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "5a7cb3168f4199ab5bdde04f0996eda1" + }, + { + "dataPath": "params_shard_72.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.18.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "885b0dff7f42261fc81f90f9282a3292" + }, + { + "dataPath": "params_shard_73.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.17.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.18.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.18.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "12e23ee7ae6dc94b8bd52ba5d6229932" + }, + { + "dataPath": "params_shard_74.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.18.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "9bbccbdb31957b918429347e45b63b5c" + }, + { + "dataPath": "params_shard_75.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.19.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "49a7dbefab06bb0104bcfad48cec1cf3" + }, + { + "dataPath": "params_shard_76.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.18.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.18.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.18.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.18.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.19.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "0ebcf2d68dce61afe922c18539674821" + }, + { + "dataPath": "params_shard_77.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "505b1a1f7074705b2de5bb72e0fc9ae2" + }, + { + "dataPath": "params_shard_78.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.19.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "8de2568c6627accebf0e3bee03837b16" + }, + { + "dataPath": "params_shard_79.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.19.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.19.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.19.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.19.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.19.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "837de53078d705d5f3a719a65f3e5053" + }, + { + "dataPath": "params_shard_80.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.20.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "93b7f6ebf67878425593eb5640fa1f82" + }, + { + "dataPath": "params_shard_81.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "844261412274ed443fb9ec2c9966c7b4" + }, + { + "dataPath": "params_shard_82.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "1d8d738b5b610dc4690f49d570a0e529" + }, + { + "dataPath": "params_shard_83.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.20.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "f839b6fe061653c3aed249226bc610c3" + }, + { + "dataPath": "params_shard_84.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.19.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.20.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.20.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.20.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.20.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.20.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "44d6cbf033dc1d562c4d1ef0c4703631" + }, + { + "dataPath": "params_shard_85.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.21.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "e77f1a9abf6d79fac87c4db199af31a1" + }, + { + "dataPath": "params_shard_86.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.21.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "f05937206787712364777a7c080be720" + }, + { + "dataPath": "params_shard_87.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.21.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "b39d6559a5b0516a03a8c2f44c481dcf" + }, + { + "dataPath": "params_shard_88.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.21.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "a6e7b6fce4242ad5de628da904937e77" + }, + { + "dataPath": "params_shard_89.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.20.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.21.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.21.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.21.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.21.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.21.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "41d323cf7f195f1d9d0f09790335df6c" + }, + { + "dataPath": "params_shard_90.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.22.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "f1b6b0c5d797874ca237532defe3f375" + }, + { + "dataPath": "params_shard_91.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "f6d40b15a1052d624f0a28ed4c696ea6" + }, + { + "dataPath": "params_shard_92.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "50c815fcdfb8e57ede661524888986fa" + }, + { + "dataPath": "params_shard_93.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.22.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "96fae934a8d60ae22b4652eb5be6feae" + }, + { + "dataPath": "params_shard_94.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.21.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.22.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.22.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.22.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.22.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.22.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "2a101991b5037cf15ed161d3c1d1cee9" + }, + { + "dataPath": "params_shard_95.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.23.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "3ae13bd0d5958b3bfce4f69da5fa8ec1" + }, + { + "dataPath": "params_shard_96.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.23.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "52313d914e2d066538891c5c772aec0e" + }, + { + "dataPath": "params_shard_97.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.22.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.23.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.23.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "5db314ba004503c0124d35108bd50f5a" + }, + { + "dataPath": "params_shard_98.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.23.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "7b523c12260fad0abe2800ba026917ac" + }, + { + "dataPath": "params_shard_99.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.24.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "d2eedee5fe61633ff3f9c0310c3232cb" + }, + { + "dataPath": "params_shard_100.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.23.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.23.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.23.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.23.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.24.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "6a2fafadd8563f0ece8ecebcaf5c7ca0" + }, + { + "dataPath": "params_shard_101.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.24.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "978c84de895aaa3db18e076ee864238f" + }, + { + "dataPath": "params_shard_102.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.24.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "6c7737caea1c6ed88813ec6de8484a57" + }, + { + "dataPath": "params_shard_103.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.24.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.24.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.24.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.24.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.24.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "36b0c2a9d46bf3eda57a7e5529ae8189" + }, + { + "dataPath": "params_shard_104.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.25.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "8d67b56a950a4204aa0db094e19e5ad9" + }, + { + "dataPath": "params_shard_105.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.25.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "aec6f7b88ae2175e9f0e6170eb10fd14" + }, + { + "dataPath": "params_shard_106.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.25.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "6b8fc124d30f9c77735fa5b3775e315f" + }, + { + "dataPath": "params_shard_107.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.25.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "175ec77711e59419eda551b9ef45562c" + }, + { + "dataPath": "params_shard_108.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.24.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.25.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.25.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.25.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.25.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.25.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "56ba099835f860efd62da457a30016e3" + }, + { + "dataPath": "params_shard_109.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.26.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "de9bbdd776fdfee439325fae921429da" + }, + { + "dataPath": "params_shard_110.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "183a1936555e6d1cac066bd63468a4c6" + }, + { + "dataPath": "params_shard_111.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "f8278c8aa0f9790a160e895c03056bf6" + }, + { + "dataPath": "params_shard_112.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.26.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "bbccec4d4249a5ff1bb4335361398d84" + }, + { + "dataPath": "params_shard_113.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.25.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.26.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.26.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.26.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.26.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.26.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "89bfa8fece476bdf8f2c530bcad217a9" + }, + { + "dataPath": "params_shard_114.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.27.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "04c7f19d975552f61f06769eacec8a71" + }, + { + "dataPath": "params_shard_115.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.27.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "1dc972e516837c5d69a02bcf0425dcf1" + }, + { + "dataPath": "params_shard_116.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.27.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "363c078354fbc0c307de26d5327d8171" + }, + { + "dataPath": "params_shard_117.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.27.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "58522760a53fca20142cc05715b724df" + }, + { + "dataPath": "params_shard_118.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.26.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.27.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.27.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.27.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.27.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.27.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "86b3240266bdb9b4dc3a08988d91aa4b" + }, + { + "dataPath": "params_shard_119.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.28.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "0e5c9d0a42c8ee67252b7c739bc661be" + }, + { + "dataPath": "params_shard_120.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.28.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "698042a86ff6925dcfc2d265a9c613b3" + }, + { + "dataPath": "params_shard_121.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.27.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.28.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.28.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "631b67f382ec9fb7a82f06a6fb532ce7" + }, + { + "dataPath": "params_shard_122.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.28.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "e9604a06f369e3ddd5ccbfdf2b52979f" + }, + { + "dataPath": "params_shard_123.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.29.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "3780e071c4cbe6ce10ddbbe7b3180400" + }, + { + "dataPath": "params_shard_124.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.28.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.28.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.28.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.28.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.29.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "d3a77ac9730cc614a5f3e35fb68b33ad" + }, + { + "dataPath": "params_shard_125.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.29.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "97989c8f7b4e732802a7430559930574" + }, + { + "dataPath": "params_shard_126.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.29.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "d1595b3066a1048e5af885e43c02fa83" + }, + { + "dataPath": "params_shard_127.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.29.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.29.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.29.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.29.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.29.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "ed06d246012d58f9cc2171c6797506d4" + }, + { + "dataPath": "params_shard_128.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.30.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "7ea20b0655c53c0e4d20a498562bd9cc" + }, + { + "dataPath": "params_shard_129.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.30.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "976c6ed45b6411dc83b2823148864b29" + }, + { + "dataPath": "params_shard_130.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.30.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "5dc5be6c9299c5293e558fe7c406a55a" + }, + { + "dataPath": "params_shard_131.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.30.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "3afd35402483fa2f58af8d7d3c4c6013" + }, + { + "dataPath": "params_shard_132.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.29.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.29.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.30.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.30.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.30.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.30.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.30.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "d750cfc7b2ed530a02d0a02c4f8d5552" + }, + { + "dataPath": "params_shard_133.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.31.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "ee0bc46d0cc64d339dd4f5e8450bcce9" + }, + { + "dataPath": "params_shard_134.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.31.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "7987093f1c56b781cf5c793378cb7a01" + }, + { + "dataPath": "params_shard_135.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.31.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "9fd7903894d3c338dfd0dea0c17a005a" + }, + { + "dataPath": "params_shard_136.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.31.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "9ae5805c0b19e7bcda71aa31b1aa7d20" + }, + { + "dataPath": "params_shard_137.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.30.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.31.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.31.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.31.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.31.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.31.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "471b8d4c2992694f2ccd5bd3b70b01cf" + }, + { + "dataPath": "params_shard_138.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.32.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "29b9098a094332b12e47122a79368346" + }, + { + "dataPath": "params_shard_139.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.32.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "20535d3d3ef8173ab7ebf44992912c64" + }, + { + "dataPath": "params_shard_140.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.32.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "3ccaee201841fc12eee8c2093bff2d5b" + }, + { + "dataPath": "params_shard_141.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.32.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "7b3d8024184d02f78471811783c26a99" + }, + { + "dataPath": "params_shard_142.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.31.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.31.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.32.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.32.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.32.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.32.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.32.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "29c26e5c96fe670e6ab2d7fbaca21892" + }, + { + "dataPath": "params_shard_143.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.33.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "27106e0bab8ccf0f08195fc8ae6ca9fa" + }, + { + "dataPath": "params_shard_144.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.33.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "de8db001a89e71c9a19713b6190bdb8e" + }, + { + "dataPath": "params_shard_145.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.32.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.32.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.33.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.33.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "ba920183df8215759247fbf3e5409bca" + }, + { + "dataPath": "params_shard_146.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.3.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "327d74802468f176e2e869b1f4a6b8f5" + }, + { + "dataPath": "params_shard_147.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.4.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "f33e337304b70406dbfa19fea555a198" + }, + { + "dataPath": "params_shard_148.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.33.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.3.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.3.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.3.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.4.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "ea539d6e6dbe48fa8ec31b3ee10df8a2" + }, + { + "dataPath": "params_shard_149.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "044c5d5ea695d24d28e6f3a4a1394fad" + }, + { + "dataPath": "params_shard_150.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.4.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "360667a75f3b45b06b5aa20026316f0e" + }, + { + "dataPath": "params_shard_151.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.4.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.4.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.4.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.4.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.4.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "aa3b198da2ac505a158e5d2b21c5ee63" + }, + { + "dataPath": "params_shard_152.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.5.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "64fcc9b40402707afcebf4eee4035fc5" + }, + { + "dataPath": "params_shard_153.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "b849dfd903d38e24540419bd518598f6" + }, + { + "dataPath": "params_shard_154.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "5f789dc553621b70fb20214e0487bb62" + }, + { + "dataPath": "params_shard_155.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.5.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "f79ed2239ca5d9a1f4607cca6246a4e5" + }, + { + "dataPath": "params_shard_156.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.4.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.5.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.5.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.5.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.5.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.5.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "e59c22862f54199aaacc1b31a152f780" + }, + { + "dataPath": "params_shard_157.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.6.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "a5a6590858adfdf098594a373c133e7a" + }, + { + "dataPath": "params_shard_158.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "938a6da5f8c4df5f7cf368804e2188b6" + }, + { + "dataPath": "params_shard_159.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "2d3001ad5c1bd8b67b1eb5892faeefcc" + }, + { + "dataPath": "params_shard_160.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.6.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "8100fc103d989eb1966ffd48683deb50" + }, + { + "dataPath": "params_shard_161.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.5.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.6.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.6.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.6.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.6.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.6.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "1323e1893efed3eba82fdd2e254a10d3" + }, + { + "dataPath": "params_shard_162.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.7.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "221623d0d9d387e6cfa3c35b5d831807" + }, + { + "dataPath": "params_shard_163.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "64bf63be17ad3324faa8b911650b653d" + }, + { + "dataPath": "params_shard_164.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "7d2e39b7fa734df9cc0318107c0e9a52" + }, + { + "dataPath": "params_shard_165.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.7.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "5eb101a230fa790775130d559a46ee73" + }, + { + "dataPath": "params_shard_166.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.6.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.7.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.7.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.7.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.7.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.7.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "9f03adb6b6c62e76d5619b7bb81a81db" + }, + { + "dataPath": "params_shard_167.bin", + "format": "raw-shard", + "nbytes": 33110016, + "records": [ + { + "name": "model.layers.7.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.8.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 14745600 + }, + { + "name": "model.layers.8.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 14759936 + } + ], + "md5sum": "416ef9e53e1916d74cfe374ed0c5e43f" + }, + { + "dataPath": "params_shard_168.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.33.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "a60c5a7502acbfc7aa1001be97394ddb" + }, + { + "dataPath": "params_shard_169.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.34.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "92869c90d04e96c354ceeb7ad6b90848" + }, + { + "dataPath": "params_shard_170.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.8.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.33.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.33.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.33.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.34.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "17e768d9a81791eba0dd4525744c96a5" + }, + { + "dataPath": "params_shard_171.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.34.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "faa8e5b9f3d8c6338b57d0165914de58" + }, + { + "dataPath": "params_shard_172.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.34.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "7d92008cbea5bac38cc991e1d8ca0fdd" + }, + { + "dataPath": "params_shard_173.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.34.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.34.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.34.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.34.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.34.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "0f4208eba5fe04b92b1d3f3d5f4ae3aa" + }, + { + "dataPath": "params_shard_174.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.35.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "0c097c9a90547d4e22b014439c6f5c59" + }, + { + "dataPath": "params_shard_175.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "557b72bafae16c7f03f709bf3a71ac07" + }, + { + "dataPath": "params_shard_176.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "6dcf7efd9b79599ea99ea3e28de38a2a" + }, + { + "dataPath": "params_shard_177.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.35.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "6acbfdd36b51ae3f33ab5fa5b24fef07" + }, + { + "dataPath": "params_shard_178.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.34.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.34.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.35.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.35.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.35.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.35.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.35.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "d9fcb75ee49a6fc46d0185b112f22465" + }, + { + "dataPath": "params_shard_179.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.36.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "839e2d682a207de98b253f09afd1ebb4" + }, + { + "dataPath": "params_shard_180.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.36.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "350160dbd348df5ebd767f179820415d" + }, + { + "dataPath": "params_shard_181.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.36.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "b4f3c875d27b97a04e79cc3cce6721a5" + }, + { + "dataPath": "params_shard_182.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.36.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "a699a059d8d0b007240bee89ada07ae0" + }, + { + "dataPath": "params_shard_183.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.35.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.36.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.36.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.36.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.36.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.36.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "a97b6698fba52ad7b4fc7f73a19824ad" + }, + { + "dataPath": "params_shard_184.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.37.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "2f287f6d4fb2a54b1afa4578995aa55b" + }, + { + "dataPath": "params_shard_185.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.37.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "a0c07d49f0425303f058e994b77c4148" + }, + { + "dataPath": "params_shard_186.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.37.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "2cd4154638cfe25e11b74425b3d1794e" + }, + { + "dataPath": "params_shard_187.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.37.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "2c95f0a229ff19cfd6981db4d1100b6a" + }, + { + "dataPath": "params_shard_188.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.36.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.36.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.37.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.37.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.37.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.37.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.37.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "4142f59d9b85ec0a922ed6fee666d142" + }, + { + "dataPath": "params_shard_189.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.38.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "05234682b6764dfffe08114c1d01dda5" + }, + { + "dataPath": "params_shard_190.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.38.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "c37b5e6de94260159944febdcad05dbb" + }, + { + "dataPath": "params_shard_191.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.37.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.37.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.38.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.38.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "eddb1742afa5ea5e27e082bcfae3e50b" + }, + { + "dataPath": "params_shard_192.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.38.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "45be5ada66f22e732c33901b2b09406d" + }, + { + "dataPath": "params_shard_193.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.39.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "b219fb4076e4993622210a6c75ebf7dc" + }, + { + "dataPath": "params_shard_194.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.38.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.38.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.38.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.38.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.38.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.38.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.39.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "41ea36c84c873c8cc25c19e31e7a5d5f" + }, + { + "dataPath": "params_shard_195.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.39.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "ddbb85d5d626796a29f73d2d7cbe2a98" + }, + { + "dataPath": "params_shard_196.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.39.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "83d343a6e4d6df0b0d70b3a6962f3216" + }, + { + "dataPath": "params_shard_197.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.39.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.39.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.39.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.39.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.39.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "1615907d4cc6bbb5ae991c44fea3ca48" + }, + { + "dataPath": "params_shard_198.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.40.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "f5827979729213b711d7ff3136a8d3e5" + }, + { + "dataPath": "params_shard_199.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.40.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "203f22bfe8d74d2cba2a168b01349671" + }, + { + "dataPath": "params_shard_200.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.40.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "a275ce43bf97e259db6489a74cd03cb6" + }, + { + "dataPath": "params_shard_201.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.40.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "9a4a270b20347f85f1698448a406c8ba" + }, + { + "dataPath": "params_shard_202.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.39.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.39.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.40.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.40.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.40.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.40.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.40.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "533cd466d7355195c9d5460a8d61f88f" + }, + { + "dataPath": "params_shard_203.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.41.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "8f030803a655ccae6ce598c735500328" + }, + { + "dataPath": "params_shard_204.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.41.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "51ecd086bbc7b03c680ed0716477a1d9" + }, + { + "dataPath": "params_shard_205.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.41.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "a07f6fb072765a0f6c539b199810a858" + }, + { + "dataPath": "params_shard_206.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.41.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "03e4b6ba356f6aeb9b106d6d4ee31169" + }, + { + "dataPath": "params_shard_207.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.40.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.40.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.41.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.41.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.41.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.41.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.41.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "958fd5266753ec0e4963f04f945c0f39" + }, + { + "dataPath": "params_shard_208.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.42.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "a0a433a0881ef311ac96e55b5acb4b43" + }, + { + "dataPath": "params_shard_209.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.42.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "7ca234091525b95e6596ba4f72b70953" + }, + { + "dataPath": "params_shard_210.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.42.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "bf3196d098404f8b44b994e24e42f3fb" + }, + { + "dataPath": "params_shard_211.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.42.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "03e483bd5c59a6bbc358276d60f9c3cf" + }, + { + "dataPath": "params_shard_212.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.41.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.41.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.42.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.42.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.42.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.42.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.42.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "4eb0ef80ce8f310a625ebd8280ed993b" + }, + { + "dataPath": "params_shard_213.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.43.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "20ad65cfcb32bdd41f67874018d2675b" + }, + { + "dataPath": "params_shard_214.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.43.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "b63f8eb97c42f0f98a34c5e88ca7c617" + }, + { + "dataPath": "params_shard_215.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.42.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.42.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.43.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.43.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "e1ddbac204646d705710ae22cf1d74bd" + }, + { + "dataPath": "params_shard_216.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.43.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "09cc0f473fa77ed7ca3d9ac909395b2a" + }, + { + "dataPath": "params_shard_217.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.44.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "c8a3b2a86be2d270688b606df59240d9" + }, + { + "dataPath": "params_shard_218.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.43.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.43.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.43.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.43.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.43.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.43.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.44.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "28fc50987dd9f5a63beb065483f0dd28" + }, + { + "dataPath": "params_shard_219.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.44.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "e378b38308537df881ddf8961d67909d" + }, + { + "dataPath": "params_shard_220.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.44.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "b952056ba3307665bf4244d9e5cbe7c9" + }, + { + "dataPath": "params_shard_221.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.44.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.44.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.44.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.44.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.44.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "11b504687dd0c8c8cf456feb24e9654a" + }, + { + "dataPath": "params_shard_222.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.45.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "c530be48df251bfcb0a52c2ee89b2083" + }, + { + "dataPath": "params_shard_223.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.45.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "aaa224e87dc489c4b7ee55f58724e3fd" + }, + { + "dataPath": "params_shard_224.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.45.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "e5d246fddfada6de3f76d667d1b120c1" + }, + { + "dataPath": "params_shard_225.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.45.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "28ff7980fe228e9b960467221b7588d3" + }, + { + "dataPath": "params_shard_226.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.44.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.44.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.45.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.45.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.45.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.45.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.45.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "8835f47c758f7adb07c5cb67936c998a" + }, + { + "dataPath": "params_shard_227.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.46.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "d45fac15620808acc9a28f08a182858b" + }, + { + "dataPath": "params_shard_228.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.46.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "3e4632a216e6ca714653466db98a5f76" + }, + { + "dataPath": "params_shard_229.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.46.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "a1c5519a4e28063322507f3a0565d058" + }, + { + "dataPath": "params_shard_230.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.46.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "c02f6bc5edd26951927162cbb5fb683d" + }, + { + "dataPath": "params_shard_231.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.45.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.45.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.46.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.46.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.46.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.46.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.46.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "29fbf8b9a8fcf5b304454583698fcfc5" + }, + { + "dataPath": "params_shard_232.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.47.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "432dbd994305841bcf32ffbed8ea2d6d" + }, + { + "dataPath": "params_shard_233.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.47.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "33c716bacb15e0b1ac0d16b440319915" + }, + { + "dataPath": "params_shard_234.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.47.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "ffefdc0b89e854f1ed8755b26463fe55" + }, + { + "dataPath": "params_shard_235.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.47.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "56421ff66d79f7baacbf5ea78acd8fd2" + }, + { + "dataPath": "params_shard_236.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.46.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.46.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.47.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.47.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.47.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.47.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.47.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "e7da40e8968ca0123448d57f27e03770" + }, + { + "dataPath": "params_shard_237.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.48.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "c1bfcf92f421aad42afcb640bd106d1e" + }, + { + "dataPath": "params_shard_238.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.48.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "cdfe3f2d70970978bd2f806e919154f1" + }, + { + "dataPath": "params_shard_239.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.47.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.47.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.48.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.48.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "6b0db1adb4ebf7fc9f148fa9ed4cb9ef" + }, + { + "dataPath": "params_shard_240.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.48.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "f1ce68cc88d67ec2f44b77128a10a294" + }, + { + "dataPath": "params_shard_241.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.49.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "574f7b8c5b81fd746516a10e7e1b8c1b" + }, + { + "dataPath": "params_shard_242.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.48.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.48.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.48.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.48.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.48.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.48.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.49.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "3d5ba2b96e0dcd4a5800404a1fd7c102" + }, + { + "dataPath": "params_shard_243.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.49.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "2cdaba74a5065d1b7df96eaa6e59b0a9" + }, + { + "dataPath": "params_shard_244.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.49.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "9dbadde1535f636a5fe98a67618b142e" + }, + { + "dataPath": "params_shard_245.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.49.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.49.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.49.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.49.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.49.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "a0e9f096c59a6db47936b9d3e136831f" + }, + { + "dataPath": "params_shard_246.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.50.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "b0a3d3d226c10d8ea9850e0849cd6c5e" + }, + { + "dataPath": "params_shard_247.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.50.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "fafe19006433cbe87d171b6f8a3f4771" + }, + { + "dataPath": "params_shard_248.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.50.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "f737835e0e7101a1f499a6f98a3b108c" + }, + { + "dataPath": "params_shard_249.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.50.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "d81453190c385865a76064244cdaeba5" + }, + { + "dataPath": "params_shard_250.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.49.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.49.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.50.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.50.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.50.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.50.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.50.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "f629bdc94ca24e10e3823b9bde176c39" + }, + { + "dataPath": "params_shard_251.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.51.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "4fcd156feebbf8a8bea7d3ff2ddd019c" + }, + { + "dataPath": "params_shard_252.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.51.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "fcbc63af8c57e276e841f3eaff7d1f48" + }, + { + "dataPath": "params_shard_253.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.51.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "4562896126b6e5f7070acc55a1b317fd" + }, + { + "dataPath": "params_shard_254.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.51.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "34513a75752d523b9a9f8bba499323fa" + }, + { + "dataPath": "params_shard_255.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.50.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.50.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.51.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.51.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.51.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.51.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.51.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "481b5c16d636d65216dedf72706213d6" + }, + { + "dataPath": "params_shard_256.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.52.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "8c97df3b6c725c6470df39b4f8c4b1f7" + }, + { + "dataPath": "params_shard_257.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.52.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "e72959bea3849d4eb9084d08cf0f0944" + }, + { + "dataPath": "params_shard_258.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.52.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "e9100d5e948965c413d82e964a3ed7c5" + }, + { + "dataPath": "params_shard_259.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.52.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "4ef711018e3b05324c77e30b59f769c1" + }, + { + "dataPath": "params_shard_260.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.51.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.51.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.52.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.52.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.52.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.52.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.52.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "730f26373aeaa811eef3434f0e145dbc" + }, + { + "dataPath": "params_shard_261.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.53.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "535009a5fea07d0c9686aa46d4e06d33" + }, + { + "dataPath": "params_shard_262.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.53.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "f0dc118b4e5f16088ecf4f417c78c3d6" + }, + { + "dataPath": "params_shard_263.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.52.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.52.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.53.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.53.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "82454b414fbf7ade99dd7912dd0807cc" + }, + { + "dataPath": "params_shard_264.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.53.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "26e8df921ca6ae48674e23ab6a9d5a1c" + }, + { + "dataPath": "params_shard_265.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.54.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "1f594b1907a6df7d086592bd696eef07" + }, + { + "dataPath": "params_shard_266.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.53.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.53.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.53.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.53.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.53.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.53.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.54.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "e0feda2dc8f4c08b2cceb5e0d59a5aa2" + }, + { + "dataPath": "params_shard_267.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.54.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "f3f875137671f8887275ca7bf884a2e0" + }, + { + "dataPath": "params_shard_268.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.54.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "8b38f92db25cc30724dbe506c5108cb6" + }, + { + "dataPath": "params_shard_269.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.54.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.54.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.54.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.54.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.54.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "83592ce9dee56a326ce7e6fa43c130fe" + }, + { + "dataPath": "params_shard_270.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.55.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "45aef1bce1d1c6d06e4785fd7fb3a65f" + }, + { + "dataPath": "params_shard_271.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.55.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "8d85df241341144f7add9cd229361753" + }, + { + "dataPath": "params_shard_272.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.55.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "4bfd0e422afb07373cf27eb4814b8f7a" + }, + { + "dataPath": "params_shard_273.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.55.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "ba20001d1d0557640f8d8c851211826a" + }, + { + "dataPath": "params_shard_274.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.54.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.54.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.55.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.55.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.55.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.55.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.55.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "706722705060142e2b147ea66d780992" + }, + { + "dataPath": "params_shard_275.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.56.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "d899c6e08a81eddf009073fe715aebca" + }, + { + "dataPath": "params_shard_276.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.56.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "d9ae02a3a70695235afadce6a7f57c52" + }, + { + "dataPath": "params_shard_277.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.56.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "2bb63e244a23e913a430d75e1987c300" + }, + { + "dataPath": "params_shard_278.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.56.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "488ca9f60a40f484a1cc5e3cef8066a1" + }, + { + "dataPath": "params_shard_279.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.55.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.55.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.56.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.56.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.56.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.56.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.56.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "a5b723f92e36ecebd8910c296da28ead" + }, + { + "dataPath": "params_shard_280.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.57.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "394a538d68d0bf72bcfb1b45a4ec0838" + }, + { + "dataPath": "params_shard_281.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.57.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "205de6325d9d38c5a3a4a22acc1bb35f" + }, + { + "dataPath": "params_shard_282.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.57.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "dc02e0c81d1e10bbf319d580fc6831a3" + }, + { + "dataPath": "params_shard_283.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.57.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "b0c832893d66678e26c82420b4d4ed39" + }, + { + "dataPath": "params_shard_284.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.56.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.56.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.57.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.57.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.57.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.57.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.57.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "35510b08db3f5673768f7a3ae79bfe31" + }, + { + "dataPath": "params_shard_285.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.58.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "bfd8e0cd62fd0f9d1e2f81c8e9fa3b52" + }, + { + "dataPath": "params_shard_286.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.58.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "c11532b68a2bf12e904a4224f33a1f49" + }, + { + "dataPath": "params_shard_287.bin", + "format": "raw-shard", + "nbytes": 32454656, + "records": [ + { + "name": "model.layers.57.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.57.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.58.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 14745600 + }, + { + "name": "model.layers.58.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 32440320 + } + ], + "md5sum": "6ec21b59cae6409b3dd9878af8c0e6e9" + }, + { + "dataPath": "params_shard_288.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.58.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "2e3469698c95ad7345e9fb3c7540263c" + }, + { + "dataPath": "params_shard_289.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.59.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "5c93411e80c815089462ec4cbd94bda9" + }, + { + "dataPath": "params_shard_290.bin", + "format": "raw-shard", + "nbytes": 25917440, + "records": [ + { + "name": "model.layers.58.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.58.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.58.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + }, + { + "name": "model.layers.58.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17039360 + }, + { + "name": "model.layers.58.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 17049600 + }, + { + "name": "model.layers.58.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25896960 + }, + { + "name": "model.layers.59.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25907200 + } + ], + "md5sum": "1f69c163a74a53345da95de2319f0529" + }, + { + "dataPath": "params_shard_291.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.59.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "b937367c2111fcc5ba4c99ce413b9754" + }, + { + "dataPath": "params_shard_292.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.59.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "f9699fb7075a93b98b12fb3d6e86158f" + }, + { + "dataPath": "params_shard_293.bin", + "format": "raw-shard", + "nbytes": 28860416, + "records": [ + { + "name": "model.layers.59.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 0 + }, + { + "name": "model.layers.59.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 8847360 + }, + { + "name": "model.layers.59.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 26542080 + }, + { + "name": "model.layers.59.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 26552320 + }, + { + "name": "model.layers.59.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 26566656 + } + ], + "md5sum": "23a4846369dfde6a806f1fbc9a84e9e2" + }, + { + "dataPath": "params_shard_294.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.60.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "c29b7ad308252e5040eb2fe2d169d26a" + }, + { + "dataPath": "params_shard_295.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.60.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "84b6caefe915cf9f21d90a6a955e928f" + }, + { + "dataPath": "params_shard_296.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.60.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "dd3c7c63e8fe53dd856e42554459a0af" + }, + { + "dataPath": "params_shard_297.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.60.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "85ed79ae981f6cd22e9383d42d309be3" + }, + { + "dataPath": "params_shard_298.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.59.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.59.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.60.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.60.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.60.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.60.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.60.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "1ffdef959858abfa93c6e1be455b9f60" + }, + { + "dataPath": "params_shard_299.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.61.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "3b044afff3b86814b8a0b4d69129119e" + }, + { + "dataPath": "params_shard_300.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.61.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "d709f4cbb8fdab0f148cfaceb6463266" + }, + { + "dataPath": "params_shard_301.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.61.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "941f93c95ad9d43051bb39cbe9ffb6de" + }, + { + "dataPath": "params_shard_302.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.61.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "6d8701e90548073e3f3e8710f8eb4083" + }, + { + "dataPath": "params_shard_303.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.60.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.60.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.61.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.61.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.61.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.61.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.61.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "70c645f2055129a239a4ab50fda87f8b" + }, + { + "dataPath": "params_shard_304.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.62.mlp.down_proj.q_weight", + "shape": [ + 5120, + 3456 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "ac5fd0cb2b1434ac499205c188abd630" + }, + { + "dataPath": "params_shard_305.bin", + "format": "raw-shard", + "nbytes": 141557760, + "records": [ + { + "name": "model.layers.62.mlp.gate_up_proj.q_weight", + "shape": [ + 55296, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 141557760, + "byteOffset": 0 + } + ], + "md5sum": "e9008621520674741da9a41f6d7eab57" + }, + { + "dataPath": "params_shard_306.bin", + "format": "raw-shard", + "nbytes": 17694720, + "records": [ + { + "name": "model.layers.62.mlp.gate_up_proj.q_scale", + "shape": [ + 55296, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 17694720, + "byteOffset": 0 + } + ], + "md5sum": "2783742a597e7e2dda5932015ea6e729" + }, + { + "dataPath": "params_shard_307.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.62.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "78ececd0dc31bf2c49cf82af258154ce" + }, + { + "dataPath": "params_shard_308.bin", + "format": "raw-shard", + "nbytes": 25921536, + "records": [ + { + "name": "model.layers.61.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.61.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.62.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14745600 + }, + { + "name": "model.layers.62.mlp.down_proj.q_scale", + "shape": [ + 5120, + 864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8847360, + "byteOffset": 14755840 + }, + { + "name": "model.layers.62.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 23603200 + }, + { + "name": "model.layers.62.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 23613440 + }, + { + "name": "model.layers.62.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 23627776 + } + ], + "md5sum": "28f5330177f894d6581f27e3c40ed9ac" + }, + { + "dataPath": "params_shard_309.bin", + "format": "raw-shard", + "nbytes": 33110016, + "records": [ + { + "name": "model.layers.62.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.62.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 13107200 + }, + { + "name": "model.layers.63.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 14745600 + }, + { + "name": "model.layers.63.self_attn.c_attn.q_weight", + "shape": [ + 7168, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 14759936 + } + ], + "md5sum": "8630ce24471ee85ca90b670ace198a75" + }, + { + "dataPath": "params_shard_310.bin", + "format": "raw-shard", + "nbytes": 17039360, + "records": [ + { + "name": "model.layers.63.self_attn.c_attn.q_scale", + "shape": [ + 7168, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2293760, + "byteOffset": 0 + }, + { + "name": "model.layers.63.self_attn.o_proj.q_weight", + "shape": [ + 5120, + 640 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 2293760 + }, + { + "name": "model.layers.63.self_attn.o_proj.q_scale", + "shape": [ + 5120, + 160 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1638400, + "byteOffset": 15400960 + } + ], + "md5sum": "30235806fff93a14ca58d06aba24486c" + } + ] +} \ No newline at end of file