diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,3577 @@ +{ + "metadata": { + "ParamSize": 311, + "ParamBytes": 796192768.0, + "BitsPerParam": 4.126114610388425 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 116686848, + "records": [ + { + "name": "model.embed_tokens.q_weight", + "shape": [ + 151936, + 192 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 116686848, + "byteOffset": 0 + } + ], + "md5sum": "56df338132ee3e4da773567d31589f8b" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 27787264, + "records": [ + { + "name": "model.embed_tokens.q_scale", + "shape": [ + 151936, + 12 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3646464, + "byteOffset": 0 + }, + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 3646464 + }, + { + "name": "model.layers.0.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 3649536 + }, + { + "name": "model.layers.0.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 10530816 + }, + { + "name": "model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 10745856 + }, + { + "name": "model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 24508416 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 24938496 + }, + { + "name": "model.layers.0.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 24941568 + }, + { + "name": "model.layers.0.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 24945664 + }, + { + "name": "model.layers.0.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 26518528 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 26567680 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 27747328 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 27784192 + } + ], + "md5sum": "776667a3add012d09e3b0d71e4170725" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 31234048, + "records": [ + { + "name": "model.layers.1.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 0 + }, + { + "name": "model.layers.1.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 6881280 + }, + { + "name": "model.layers.1.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 7096320 + }, + { + "name": "model.layers.1.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 20858880 + }, + { + "name": "model.layers.1.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 21288960 + }, + { + "name": "model.layers.1.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 21292032 + }, + { + "name": "model.layers.1.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 21296128 + }, + { + "name": "model.layers.1.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 22868992 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 22918144 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 24097792 + }, + { + "name": "model.layers.10.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 24134656 + }, + { + "name": "model.layers.10.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 24137728 + }, + { + "name": "model.layers.10.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 31019008 + } + ], + "md5sum": "8916a08da2d8072077fe8717690385f2" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.10.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.10.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.10.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.10.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.10.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.10.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.11.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.11.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.11.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "f9e7d3c52a9a86e477a2fe74091b5aab" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.11.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.11.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.11.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.11.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.11.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.12.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.12.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.12.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "4c44fdc389bf8e2bd2b61ee00c6127ad" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.12.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.12.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.12.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.12.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.12.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.12.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.13.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.13.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.13.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "1af26be55dc75f7b2763e380926ba56f" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.13.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.13.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.13.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.13.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.13.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.14.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.14.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.14.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "cbd17976cc3d4785adba9889b80f8ffc" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.14.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.14.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.14.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.14.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.14.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.14.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.15.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.15.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.15.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "e34378f338e09f625c347f0b12b45b4c" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.15.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.15.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.15.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.15.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.15.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.16.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.16.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.16.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "85c08670b1376dafa71d391af0a3a0af" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.16.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.16.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.16.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.16.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.16.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.16.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.17.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.17.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.17.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "d73ee9df08971663bc3df4e94cf52cfe" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.17.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.17.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.17.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.17.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.17.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.18.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.18.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.18.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "ae8a9e01c30cee60f8e838dd3f278c98" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.18.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.18.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.18.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.18.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.18.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.18.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.19.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.19.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.19.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "0b84daf1ecbc8dcd25681dc6f33caa79" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.19.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.19.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.19.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.19.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.19.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.2.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.2.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.2.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "4659f514968f8fb437e673bab6eb5894" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.2.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.2.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.2.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.2.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.2.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.20.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.20.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.20.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "c1e3f9e3b53e25a20cc964d31eb607bb" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.20.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.20.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.20.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.20.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.20.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.21.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.21.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.21.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "46c1c033990e01e1f8348e2a656bf4e2" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.21.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.21.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.21.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.21.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.21.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.21.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.22.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.22.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.22.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "0a2778bc53dfce0d6ab8ef62ec996e5a" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.22.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.22.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.22.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.22.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.22.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.23.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.23.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.23.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "9428a4c316b7e27fe23d3ec6ad7206e2" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.23.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.23.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.23.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.23.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.23.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.23.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.24.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.24.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.24.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "368f39cb5da018ca3214656488c57352" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.24.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.24.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.24.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.24.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.24.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.24.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.25.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.25.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.25.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "3a95df9f2c42604de34534bcc153eb23" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.25.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.25.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.25.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.25.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.25.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.25.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.26.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.26.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.26.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "63566d5258eb3d9daa9c5e842d8cf858" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.26.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.26.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.26.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.26.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.26.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.27.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.27.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.27.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "ed6acf51fe05f24ac531f890a647aac5" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.27.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.27.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.27.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.27.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.27.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.27.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.3.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.3.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.3.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "6ff7975ea0d70c3e3124493fe7c2644b" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.3.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.3.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.3.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.3.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.3.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.4.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.4.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.4.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "b79c459d51aecf558a7799314a0aa841" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.4.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.4.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.4.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.4.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.4.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.5.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.5.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.5.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "9cd59be75c14befba350003fc904c8cd" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.5.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.5.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.5.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.5.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.5.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.6.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.6.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.6.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "42acc4d2cc1db9ba3809ea5460a6dbf3" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.6.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.6.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.6.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.6.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.6.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.7.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.7.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.7.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "35c8f796a3ec4f1e1453ee94c82f0cbe" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.7.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.7.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.7.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.7.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.7.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.8.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.8.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.8.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "f1b4e3503b0d8bb1115fcd6c5881ad88" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 24137728, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.8.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.8.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.8.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.8.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.8.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.layers.9.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + }, + { + "name": "model.layers.9.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 17041408 + }, + { + "name": "model.layers.9.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 23922688 + } + ], + "md5sum": "c438e4aa32bc9e200a389f4e1dbe2514" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 17041408, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 0 + }, + { + "name": "model.layers.9.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 13762560 + }, + { + "name": "model.layers.9.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 14192640 + }, + { + "name": "model.layers.9.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 14195712 + }, + { + "name": "model.layers.9.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 14199808 + }, + { + "name": "model.layers.9.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 15772672 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 15821824 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 17001472 + }, + { + "name": "model.norm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 17038336 + } + ], + "md5sum": "873891ec0cc91617ec05d009e4aeb7bd" + } + ] +} \ No newline at end of file