diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,7103 @@ +{ + "metadata": { + "ParamSize": 533, + "ParamBytes": 7617046528.0, + "BitsPerParam": 4.125675919921857 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 389283840, + "records": [ + { + "name": "lm_head.q_weight", + "shape": [ + 640, + 152064 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 389283840, + "byteOffset": 0 + } + ], + "md5sum": "20e1eaa810284bbb6b4b85d379d21b55" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.47.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "8e41416cbc132e6bac0d3bfa61ada674" + }, + { + "dataPath": "params_shard_2.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": "8bada77a50e129989cde4031e8180434" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.0.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "ed6275c5e40776cbdd0052b3cc5a3343" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "aec90e01b7fc05e3bb29573106be4a3d" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.0.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "b525bb3fc24c479f0daf31416c39048c" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 29372416, + "records": [ + { + "name": "lm_head.q_scale", + "shape": [ + 40, + 152064 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 12165120, + "byteOffset": 0 + }, + { + "name": "model.layers.47.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 12165120 + }, + { + "name": "model.norm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13271040 + }, + { + "name": "model.embed_tokens.q_scale", + "shape": [ + 152064, + 40 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 12165120, + "byteOffset": 13281280 + }, + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 25446400 + }, + { + "name": "model.layers.0.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 25456640 + }, + { + "name": "model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 26562560 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 28774400 + }, + { + "name": "model.layers.0.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 28784640 + }, + { + "name": "model.layers.0.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 28798976 + } + ], + "md5sum": "b83be1250bbb2ddba0dc6981f250bc1e" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.1.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "24daf439c5f45b3ed1ea24f555fb1244" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.1.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "3b3930a94aac3b4de2100c915f033436" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.1.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "10bc94f7c27b9a9270aa6dfb59fd31dc" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.2.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "3496db47f00cbe001d5bc1558e7ec428" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "ba28fc97857fb41aed933468e74741fc" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.1.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.1.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.1.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.1.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.1.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.2.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.2.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "b727669bd6e1b8d1b9f6d0e3492c2a1f" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.2.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.2.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.2.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.2.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "9176846749b78ac5adcccf44a297d22f" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.3.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "6cbf465d1fde2958047b870de632dd90" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "bc320e12d2a80d7bfcb8347697715869" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.3.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "57115a178eb7d6fa47438afa1a02df58" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.4.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "43efc508141dbb5d7cb7c585f83a71aa" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 31567872, + "records": [ + { + "name": "model.layers.2.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.3.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.3.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.3.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.3.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.3.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.3.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.4.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.4.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30969856 + }, + { + "name": "model.layers.4.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 30980096 + }, + { + "name": "model.layers.4.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 30994432 + } + ], + "md5sum": "1a4dec7d92f824113aba17d511e02668" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.10.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "4c223c5867b9b797691ea3a303227cc9" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.10.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "9c89531a7fb2f60f4e1be76fa21c0f3c" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.10.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "e785af5269257b7445cd3858f5192b4d" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "962c66b74d434c1bf5b46ae76fd8b2b9" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.11.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "ff0767c01477a621c3279db92093865e" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 33206272, + "records": [ + { + "name": "model.layers.4.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.10.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.10.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.10.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.10.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.10.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.10.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.11.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.11.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 30969856 + }, + { + "name": "model.layers.11.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 33181696 + }, + { + "name": "model.layers.11.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 33191936 + } + ], + "md5sum": "8788cd2fc0ff7c0542c9dab47f9e2633" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.4.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "89f19161f316a3b086055834891343cc" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "def95b32af10158ccdcdb5e956eca61b" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.5.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "778bcf9a60ce6b08e53bd7e32e6fa0ec" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "a2707f25e926f70f5491cc9b27d472d7" + }, + { + "dataPath": "params_shard_29.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.5.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "f19ff8b7e26bbae6279582ddadd4b45e" + }, + { + "dataPath": "params_shard_30.bin", + "format": "raw-shard", + "nbytes": 21334016, + "records": [ + { + "name": "model.layers.11.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 0 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 573440 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13680640 + }, + { + "name": "model.layers.4.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 14090240 + }, + { + "name": "model.layers.4.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 15196160 + }, + { + "name": "model.layers.5.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17408000 + }, + { + "name": "model.layers.5.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 17418240 + }, + { + "name": "model.layers.5.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 18524160 + }, + { + "name": "model.layers.5.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 20736000 + }, + { + "name": "model.layers.5.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 20746240 + }, + { + "name": "model.layers.5.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20760576 + } + ], + "md5sum": "22c7456eefa49bdf4a86f5436ea5eebd" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.6.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "e3f272f1a1c7e4a14d529f1936429943" + }, + { + "dataPath": "params_shard_32.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "e65366ab9166d9ec2b3323237b6adeaf" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.6.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "7ed272cc7847f37a06147e8da1557865" + }, + { + "dataPath": "params_shard_34.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.7.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "6a66bbd5a31764ffd20be4f782abd5ad" + }, + { + "dataPath": "params_shard_35.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "68379f8ddcf20a04c973bba32548ddc5" + }, + { + "dataPath": "params_shard_36.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.5.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.6.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.6.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.6.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.6.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.6.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.6.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.7.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.7.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "fbb97dbee8bb0f80696bc6e2a656a704" + }, + { + "dataPath": "params_shard_37.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.7.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.7.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.7.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.7.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "8732902ee76dcd6677eb29af24d545b9" + }, + { + "dataPath": "params_shard_38.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.8.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "147797db058f88d88b11bcba71e1d56e" + }, + { + "dataPath": "params_shard_39.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "3b02cb97d9f639d7f3b71c882f3d73ee" + }, + { + "dataPath": "params_shard_40.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.8.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "90aab4e7e272d64f07fe0f8c17566dc0" + }, + { + "dataPath": "params_shard_41.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.9.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "517ebc9854ae5b17d37a1bfe936bf26e" + }, + { + "dataPath": "params_shard_42.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "a405d3a6820126b6141525f6eaaa935e" + }, + { + "dataPath": "params_shard_43.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.7.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.8.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.8.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.8.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.8.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.8.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.8.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.9.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.9.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "97c1cad8577ba1ea717706a6bd321bcf" + }, + { + "dataPath": "params_shard_44.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.9.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.9.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.9.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.9.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "883b731c2c00851cc215fb1de2e3e952" + }, + { + "dataPath": "params_shard_45.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.11.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "3da296611ceb2658a1398b17002a5f77" + }, + { + "dataPath": "params_shard_46.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.12.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "4541939633caf1a7dfb050f393a81e68" + }, + { + "dataPath": "params_shard_47.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.12.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "9bbadef6b04c05bb914529feb350450c" + }, + { + "dataPath": "params_shard_48.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.12.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "23b066cff7a5433594f80c383812c750" + }, + { + "dataPath": "params_shard_49.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.13.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "334e091b393f70af1b30001a4e16e34c" + }, + { + "dataPath": "params_shard_50.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "3e687c326fe7a5b8abe77fe655264558" + }, + { + "dataPath": "params_shard_51.bin", + "format": "raw-shard", + "nbytes": 33181696, + "records": [ + { + "name": "model.layers.9.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.11.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13516800 + }, + { + "name": "model.layers.12.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 14622720 + }, + { + "name": "model.layers.12.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 14632960 + }, + { + "name": "model.layers.12.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 15738880 + }, + { + "name": "model.layers.12.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 17950720 + }, + { + "name": "model.layers.12.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 17960960 + }, + { + "name": "model.layers.12.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 17975296 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 18548736 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 31655936 + }, + { + "name": "model.layers.13.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 32065536 + }, + { + "name": "model.layers.13.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 32075776 + } + ], + "md5sum": "95c89850fd802bebc6841e08972e0601" + }, + { + "dataPath": "params_shard_52.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.13.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.13.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.13.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.13.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "157df6d68c2f7040ad85b528623f93ae" + }, + { + "dataPath": "params_shard_53.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.14.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "e476ab162b9c83f931fc0a6eebd5e4a4" + }, + { + "dataPath": "params_shard_54.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.14.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "7e398e1f7b811238532104c9b484778b" + }, + { + "dataPath": "params_shard_55.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.14.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "076d5ff9ae1ae35c68f4a0b59d450882" + }, + { + "dataPath": "params_shard_56.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.15.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "4f287248f9ad60e71bba3694be7c50ed" + }, + { + "dataPath": "params_shard_57.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "a8c7134cfc092a10f76af48c2029c038" + }, + { + "dataPath": "params_shard_58.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.13.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.14.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.14.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.14.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.14.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.14.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.14.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.15.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.15.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "8ba7310e358f85ebaefd53a08ec9badc" + }, + { + "dataPath": "params_shard_59.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.15.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.15.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.15.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.15.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "279e828225a3c0f11ddd1644357fda53" + }, + { + "dataPath": "params_shard_60.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.16.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "ad1ef5668e129ffffe401f6d72e0e1af" + }, + { + "dataPath": "params_shard_61.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.16.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "cb93de571be0a62c7e27ba4d13250873" + }, + { + "dataPath": "params_shard_62.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.16.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "1045e09ecac923fa43d279dd2a9a9a87" + }, + { + "dataPath": "params_shard_63.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.17.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "d081992438bc4193e7423064429b5fe3" + }, + { + "dataPath": "params_shard_64.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "fb0702109b365b888ae09ec55dcdcc03" + }, + { + "dataPath": "params_shard_65.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.15.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.16.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.16.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.16.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.16.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.16.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.16.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.17.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.17.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "9250657d70d4d85051e557a8195eee04" + }, + { + "dataPath": "params_shard_66.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.17.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.17.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.17.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.17.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "f99c9db064eeb28b7eb88f99e46c5b30" + }, + { + "dataPath": "params_shard_67.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.18.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "4bc28a812d85c4e052ec9ed26825869c" + }, + { + "dataPath": "params_shard_68.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.18.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "08fc6534768e6dc33e155265343c5e34" + }, + { + "dataPath": "params_shard_69.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.18.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "fbaf71230b33fba1aa04f8f4e4bd1cb0" + }, + { + "dataPath": "params_shard_70.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.19.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "f99966358ec57391494005a78138fa24" + }, + { + "dataPath": "params_shard_71.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "00dcd3b61b067153c8c48f07903d6c3d" + }, + { + "dataPath": "params_shard_72.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.17.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.18.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.18.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 13527040 + }, + { + "name": "model.layers.18.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 15738880 + }, + { + "name": "model.layers.18.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 15749120 + }, + { + "name": "model.layers.18.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 15763456 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 16336896 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 29444096 + }, + { + "name": "model.layers.18.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 29853696 + }, + { + "name": "model.layers.19.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.19.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "f6a1b22a9b0ccb6205c9c483811392fd" + }, + { + "dataPath": "params_shard_73.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.19.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.19.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.19.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.19.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "2fb73dff231c8f81cc8eb024f26ef72f" + }, + { + "dataPath": "params_shard_74.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.20.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "87b20c73babfadbd13f648865a2b7b1a" + }, + { + "dataPath": "params_shard_75.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "081836a524de4ed37821a782a6620535" + }, + { + "dataPath": "params_shard_76.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.20.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "63db604a69ef379e23dd3e06c54721d7" + }, + { + "dataPath": "params_shard_77.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.21.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "23f36b8145641cd3f2ed5ec908f8c98b" + }, + { + "dataPath": "params_shard_78.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.21.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "407f78ebbe7670798832e63eb11667fb" + }, + { + "dataPath": "params_shard_79.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.19.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.20.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.20.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.20.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.20.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.20.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.20.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.21.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.21.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "9c91841922ac33002a28d1f873783fac" + }, + { + "dataPath": "params_shard_80.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.21.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.21.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.21.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.21.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.21.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "dcaba4611028874b0e90c9d44182d31c" + }, + { + "dataPath": "params_shard_81.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.22.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "ded491eb7733fa5a6d38046e012df41f" + }, + { + "dataPath": "params_shard_82.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "87bbc72a4ca55b91aaa75609c3f748da" + }, + { + "dataPath": "params_shard_83.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.22.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "d0e276a963c759e51b7f977f13eb59fa" + }, + { + "dataPath": "params_shard_84.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.23.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "c246de1fd3f29e8e52c5cc97c60d76f0" + }, + { + "dataPath": "params_shard_85.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.23.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "101b80c79be5f444300294b265eb00e4" + }, + { + "dataPath": "params_shard_86.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.21.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.22.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.22.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.22.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.22.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.22.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.22.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.23.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.23.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "4a1b38a73b9200357ec132e5bcfdcb72" + }, + { + "dataPath": "params_shard_87.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.23.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.23.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.23.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.23.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.23.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "499c860ef8518d03b0217c88aae69755" + }, + { + "dataPath": "params_shard_88.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.24.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "3a21ef762b806c72971699ace64c45a1" + }, + { + "dataPath": "params_shard_89.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.24.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "221084f4224e6bb0c25fa2b0f9af849e" + }, + { + "dataPath": "params_shard_90.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.24.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "91aec934a6395a79f56c7bd543831320" + }, + { + "dataPath": "params_shard_91.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.25.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "441fcf4c2eae4d5bf29e88816d9b9a9a" + }, + { + "dataPath": "params_shard_92.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.25.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "55665fd2865ec35e8195d2c9f7b0ce5a" + }, + { + "dataPath": "params_shard_93.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.23.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.24.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.24.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.24.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.24.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.24.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.24.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.25.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.25.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "a8c44aa0d05a3fd38cb733905705e335" + }, + { + "dataPath": "params_shard_94.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.25.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.25.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.25.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.25.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.25.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "443098c9b07bc2aa358c763aca81d70b" + }, + { + "dataPath": "params_shard_95.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.26.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "79793b9e65d680d3bdf40391107782e4" + }, + { + "dataPath": "params_shard_96.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "b691922fcbdd725dbb74d81ac9062497" + }, + { + "dataPath": "params_shard_97.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.26.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "a981c72fd891f370975786c6d2655c01" + }, + { + "dataPath": "params_shard_98.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.27.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "06e7e82bb5bacb702cc7ba71056cbb2b" + }, + { + "dataPath": "params_shard_99.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.27.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "d84c06bf8e10705b38f955417851fe70" + }, + { + "dataPath": "params_shard_100.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.25.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.26.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.26.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.26.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.26.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.26.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.26.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.27.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.27.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "20618374b597136239e10d7a89570b18" + }, + { + "dataPath": "params_shard_101.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.27.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.27.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.27.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.27.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.27.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "8800062f10da70d5ce1d0b420315e783" + }, + { + "dataPath": "params_shard_102.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.28.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "2f976fa7e96d5fac4452b5565f72e95e" + }, + { + "dataPath": "params_shard_103.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.28.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "1827e86bcc4679737f4ed98bc671ebe7" + }, + { + "dataPath": "params_shard_104.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.28.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "e4282a1398cefdf968649a31bbf6aa52" + }, + { + "dataPath": "params_shard_105.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.29.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "9779a6a98cdf6ce4dd8f69c02c05011b" + }, + { + "dataPath": "params_shard_106.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.29.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "b6f3b989c1907c873ff45f79979541d7" + }, + { + "dataPath": "params_shard_107.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.27.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.28.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.28.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.28.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.28.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.28.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.28.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.29.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.29.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "22fc4e29b16f4655d5af0cac3a217fe0" + }, + { + "dataPath": "params_shard_108.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.29.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.29.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.29.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.29.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.29.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "0c3b4c1c9ca81e4bcda322638b28007e" + }, + { + "dataPath": "params_shard_109.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.30.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "25392de55c529613b21bec9e42f38e0b" + }, + { + "dataPath": "params_shard_110.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.30.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "be8de9b2e07b5fc6986924445edb241b" + }, + { + "dataPath": "params_shard_111.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.30.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "b4e253d2dbaeb0e5902afd36424f8ed3" + }, + { + "dataPath": "params_shard_112.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.31.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "ca70fe8890c604882e78aed48b6d3740" + }, + { + "dataPath": "params_shard_113.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.31.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "863c9b6b6a750f00f0c720e81d79a469" + }, + { + "dataPath": "params_shard_114.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.29.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.29.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.30.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.30.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.30.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.30.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.30.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.30.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.31.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.31.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "9995a8ea7186bb097350824709fee66e" + }, + { + "dataPath": "params_shard_115.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.31.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.31.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.31.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.31.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.31.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "b073095b99ed063a9002456d2c555d81" + }, + { + "dataPath": "params_shard_116.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.32.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "5a8fcaed3ff7f1cc9ff2a16383c88f27" + }, + { + "dataPath": "params_shard_117.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.32.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "b2bd268a7f00d1679078f0463e9cafc9" + }, + { + "dataPath": "params_shard_118.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.32.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "e008bc0314c9d6476dca5c8fe8ab0036" + }, + { + "dataPath": "params_shard_119.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.33.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "19991cc49f46a8a509506020909921c0" + }, + { + "dataPath": "params_shard_120.bin", + "format": "raw-shard", + "nbytes": 31567872, + "records": [ + { + "name": "model.layers.31.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.31.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.32.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.32.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.32.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.32.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.32.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.32.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.32.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.32.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.33.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.33.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30969856 + }, + { + "name": "model.layers.33.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 30980096 + }, + { + "name": "model.layers.33.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 30994432 + } + ], + "md5sum": "6e3dbf9ed492580ddc1fa84c7fda8d9f" + }, + { + "dataPath": "params_shard_121.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.33.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "5770aa40d0abcf11d40a76e1daa8915c" + }, + { + "dataPath": "params_shard_122.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.33.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "0557518cd1983ee1f824638c4ce28b02" + }, + { + "dataPath": "params_shard_123.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.34.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "4e0bad3cb40dc7be6ea09a7050ca5c2f" + }, + { + "dataPath": "params_shard_124.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.34.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "c086c38a078a13da61ae566db13b4a5d" + }, + { + "dataPath": "params_shard_125.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.34.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "a42a2659ceb71d58b9d1d2a2600d68ed" + }, + { + "dataPath": "params_shard_126.bin", + "format": "raw-shard", + "nbytes": 20760576, + "records": [ + { + "name": "model.layers.33.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.33.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13516800 + }, + { + "name": "model.layers.33.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14622720 + }, + { + "name": "model.layers.34.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16834560 + }, + { + "name": "model.layers.34.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 16844800 + }, + { + "name": "model.layers.34.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 17950720 + }, + { + "name": "model.layers.34.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 20162560 + }, + { + "name": "model.layers.34.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 20172800 + }, + { + "name": "model.layers.34.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20187136 + } + ], + "md5sum": "24e67af860318a291a558fcdb6bfe56e" + }, + { + "dataPath": "params_shard_127.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.35.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "4844b7dc80a55fdd60e498c1a629f327" + }, + { + "dataPath": "params_shard_128.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "c49829bd753c72c2a2769a8fd40ac922" + }, + { + "dataPath": "params_shard_129.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.35.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "842219cf4345d4209f959394603108e3" + }, + { + "dataPath": "params_shard_130.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.36.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "1701425f65e5001b74f1eba67faaee1b" + }, + { + "dataPath": "params_shard_131.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.36.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "c7fe36c8a8b41d760a6801c92b4137f8" + }, + { + "dataPath": "params_shard_132.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.34.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.34.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.35.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.35.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.35.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.35.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.35.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.35.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.36.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.36.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "c7ad17fe86c5e743ba0060360c9a23b0" + }, + { + "dataPath": "params_shard_133.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.36.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.36.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.36.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.36.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.36.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "20d445a077b85a40c38da5a082310c89" + }, + { + "dataPath": "params_shard_134.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.37.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "bb8b09e73036f2e033d5c8e5820f8709" + }, + { + "dataPath": "params_shard_135.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.37.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "24444ee4db30bb109d9ee894b6db51a8" + }, + { + "dataPath": "params_shard_136.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.37.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "55744e3bc353e86e26d746a92a2ae967" + }, + { + "dataPath": "params_shard_137.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.38.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "5aa9bade93003774f07d48d3620bcdf2" + }, + { + "dataPath": "params_shard_138.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.38.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "fd565326f0023fc7bbefa23e059fb827" + }, + { + "dataPath": "params_shard_139.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.36.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.36.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.37.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.37.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.37.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.37.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.37.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.37.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.37.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.37.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.38.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.38.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "d716313ff6013e8750f84057b6a5c687" + }, + { + "dataPath": "params_shard_140.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.38.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.38.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.38.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.38.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.38.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "c91eac02c72e4507b0c0d0bce81e9246" + }, + { + "dataPath": "params_shard_141.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.39.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "9ebc4b8d0f70ca784a6200033a8904a8" + }, + { + "dataPath": "params_shard_142.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.39.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "6f1d78843b597e1a023c23a2fcf888d1" + }, + { + "dataPath": "params_shard_143.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.39.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "8a00efc38a8c51256ff190501c1820c9" + }, + { + "dataPath": "params_shard_144.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.40.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "91adb0d18b953c5354330ad14620a505" + }, + { + "dataPath": "params_shard_145.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.40.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "56825d882c609482effb3cc7a7e19445" + }, + { + "dataPath": "params_shard_146.bin", + "format": "raw-shard", + "nbytes": 33206272, + "records": [ + { + "name": "model.layers.38.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.38.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.39.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.39.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.39.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.39.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.39.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.39.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.39.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.39.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.40.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.40.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 30969856 + }, + { + "name": "model.layers.40.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 33181696 + }, + { + "name": "model.layers.40.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 33191936 + } + ], + "md5sum": "708fe6cfcfb7571afe45a85e55a96dc8" + }, + { + "dataPath": "params_shard_147.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.40.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "88414932884b7f6242f7c007abdb1596" + }, + { + "dataPath": "params_shard_148.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.41.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "8ef3446828945ca1c68fad3730ad95de" + }, + { + "dataPath": "params_shard_149.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.41.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "5f1021bab2b3f108c3dcde7ff53c092a" + }, + { + "dataPath": "params_shard_150.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.41.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "a6ef389f728608655e44aa1d7219e226" + }, + { + "dataPath": "params_shard_151.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.42.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "86a0b8ec91e4923ef9b92a239a87adfc" + }, + { + "dataPath": "params_shard_152.bin", + "format": "raw-shard", + "nbytes": 32649216, + "records": [ + { + "name": "model.layers.40.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 0 + }, + { + "name": "model.layers.40.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 573440 + }, + { + "name": "model.layers.40.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13680640 + }, + { + "name": "model.layers.40.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 14090240 + }, + { + "name": "model.layers.41.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 15196160 + }, + { + "name": "model.layers.41.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 15206400 + }, + { + "name": "model.layers.41.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 16312320 + }, + { + "name": "model.layers.41.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 18524160 + }, + { + "name": "model.layers.41.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 18534400 + }, + { + "name": "model.layers.41.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 18548736 + }, + { + "name": "model.layers.41.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 19122176 + }, + { + "name": "model.layers.41.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 32229376 + }, + { + "name": "model.layers.42.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 32638976 + } + ], + "md5sum": "edc7346974a77eb985df62724be54715" + }, + { + "dataPath": "params_shard_153.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.42.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "a3a1e4f8273f18e4d58cf897f960b3bc" + }, + { + "dataPath": "params_shard_154.bin", + "format": "raw-shard", + "nbytes": 22265856, + "records": [ + { + "name": "model.layers.42.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 0 + }, + { + "name": "model.layers.42.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 1105920 + }, + { + "name": "model.layers.42.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 3317760 + }, + { + "name": "model.layers.42.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 3328000 + }, + { + "name": "model.layers.42.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 3342336 + }, + { + "name": "model.layers.42.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 21692416 + } + ], + "md5sum": "969987530f4a2e48d330a4b06c775f26" + }, + { + "dataPath": "params_shard_155.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.43.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "b6ae61e50c261d7d2baf3429b5c3c329" + }, + { + "dataPath": "params_shard_156.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.43.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "50293c945b5f019d4bac3f5cee552cf5" + }, + { + "dataPath": "params_shard_157.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.43.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "2818982547a8b89e83203a64252101ec" + }, + { + "dataPath": "params_shard_158.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.44.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "4e32f63d39b1953a58ad25a6d053b6d5" + }, + { + "dataPath": "params_shard_159.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.44.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "06afee7c5b5d2f46e11cfd2770d6d121" + }, + { + "dataPath": "params_shard_160.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.42.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.42.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.43.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.43.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.43.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.43.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.43.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.43.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.43.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.43.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.44.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.44.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "2cef76d95b0bd82e8b25e5917460629f" + }, + { + "dataPath": "params_shard_161.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.44.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.44.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.44.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.44.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.44.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "e0734f03aecc2d34e33073639f57b411" + }, + { + "dataPath": "params_shard_162.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.45.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "b9ea5c5665ca833951601cecd7a954ea" + }, + { + "dataPath": "params_shard_163.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.45.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "5dbbe5558e7de476dedbfd12e183b2bb" + }, + { + "dataPath": "params_shard_164.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.45.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "5e8e6772dbe9835a94d198c749f66da2" + }, + { + "dataPath": "params_shard_165.bin", + "format": "raw-shard", + "nbytes": 35389440, + "records": [ + { + "name": "model.layers.46.mlp.down_proj.q_weight", + "shape": [ + 1728, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 35389440, + "byteOffset": 0 + } + ], + "md5sum": "9ccc252aecc59f245d4c22da2c2d2196" + }, + { + "dataPath": "params_shard_166.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.46.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "2ec4da048ac75847dfec61a07f993e0d" + }, + { + "dataPath": "params_shard_167.bin", + "format": "raw-shard", + "nbytes": 32075776, + "records": [ + { + "name": "model.layers.44.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.44.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.45.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.45.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 13527040 + }, + { + "name": "model.layers.45.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 14632960 + }, + { + "name": "model.layers.45.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 16844800 + }, + { + "name": "model.layers.45.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 16855040 + }, + { + "name": "model.layers.45.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 16869376 + }, + { + "name": "model.layers.45.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 17442816 + }, + { + "name": "model.layers.45.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 30550016 + }, + { + "name": "model.layers.46.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 30959616 + }, + { + "name": "model.layers.46.mlp.down_proj.q_scale", + "shape": [ + 108, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1105920, + "byteOffset": 30969856 + } + ], + "md5sum": "64e07e3f9ae8eb219eb026714a239185" + }, + { + "dataPath": "params_shard_168.bin", + "format": "raw-shard", + "nbytes": 21159936, + "records": [ + { + "name": "model.layers.46.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 0 + }, + { + "name": "model.layers.46.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 2211840 + }, + { + "name": "model.layers.46.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 2222080 + }, + { + "name": "model.layers.46.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 2236416 + }, + { + "name": "model.layers.46.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 20586496 + } + ], + "md5sum": "a831a9ba62e1ee97bed7c0d227550934" + }, + { + "dataPath": "params_shard_169.bin", + "format": "raw-shard", + "nbytes": 70778880, + "records": [ + { + "name": "model.layers.47.mlp.gate_up_proj.q_weight", + "shape": [ + 640, + 27648 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 70778880, + "byteOffset": 0 + } + ], + "md5sum": "ee7d26a2e1d3dccb1c09cc79727f018a" + }, + { + "dataPath": "params_shard_170.bin", + "format": "raw-shard", + "nbytes": 18350080, + "records": [ + { + "name": "model.layers.47.self_attn.c_attn.q_weight", + "shape": [ + 640, + 7168 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 18350080, + "byteOffset": 0 + } + ], + "md5sum": "ce6899efff51dc27d81d0d86cd7d0142" + }, + { + "dataPath": "params_shard_171.bin", + "format": "raw-shard", + "nbytes": 29853696, + "records": [ + { + "name": "model.layers.46.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 0 + }, + { + "name": "model.layers.46.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 13107200 + }, + { + "name": "model.layers.47.input_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 13516800 + }, + { + "name": "model.layers.47.mlp.gate_up_proj.q_scale", + "shape": [ + 40, + 27648 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2211840, + "byteOffset": 13527040 + }, + { + "name": "model.layers.47.post_attention_layernorm.weight", + "shape": [ + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10240, + "byteOffset": 15738880 + }, + { + "name": "model.layers.47.self_attn.c_attn.bias", + "shape": [ + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336, + "byteOffset": 15749120 + }, + { + "name": "model.layers.47.self_attn.c_attn.q_scale", + "shape": [ + 40, + 7168 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 573440, + "byteOffset": 15763456 + }, + { + "name": "model.layers.47.self_attn.o_proj.q_weight", + "shape": [ + 640, + 5120 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13107200, + "byteOffset": 16336896 + }, + { + "name": "model.layers.47.self_attn.o_proj.q_scale", + "shape": [ + 40, + 5120 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 409600, + "byteOffset": 29444096 + } + ], + "md5sum": "8e20de3a343eb0af66fd22ed440dd731" + } + ] +} \ No newline at end of file