diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,3607 @@ +{ + "metadata": { + "ParamSize": 313, + "ParamBytes": 916526080.0, + "BitsPerParam": 4.125968235675442 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 116686848, + "records": [ + { + "name": "lm_head.q_weight", + "shape": [ + 192, + 151936 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 116686848, + "byteOffset": 0 + } + ], + "md5sum": "76199e1a33a626c1efea6dd9dca0ce85" + }, + { + "dataPath": "params_shard_1.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": "929843bc68aad3eef8fed3279432ee86" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 31433728, + "records": [ + { + "name": "lm_head.q_scale", + "shape": [ + 12, + 151936 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3646464, + "byteOffset": 0 + }, + { + "name": "model.embed_tokens.q_scale", + "shape": [ + 151936, + 12 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3646464, + "byteOffset": 3646464 + }, + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 7292928 + }, + { + "name": "model.layers.0.mlp.down_proj.q_weight", + "shape": [ + 1120, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6881280, + "byteOffset": 7296000 + }, + { + "name": "model.layers.0.mlp.down_proj.q_scale", + "shape": [ + 70, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 215040, + "byteOffset": 14177280 + }, + { + "name": "model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 192, + 17920 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13762560, + "byteOffset": 14392320 + }, + { + "name": "model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 12, + 17920 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 430080, + "byteOffset": 28154880 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 28584960 + }, + { + "name": "model.layers.0.self_attn.c_attn.bias", + "shape": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 28588032 + }, + { + "name": "model.layers.0.self_attn.c_attn.q_weight", + "shape": [ + 192, + 2048 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28592128 + }, + { + "name": "model.layers.0.self_attn.c_attn.q_scale", + "shape": [ + 12, + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 49152, + "byteOffset": 30164992 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 192, + 1536 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 1179648, + "byteOffset": 30214144 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 12, + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36864, + "byteOffset": 31393792 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 1536 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3072, + "byteOffset": 31430656 + } + ], + "md5sum": "0ab034f191592452e57d20cd2d8ef681" + }, + { + "dataPath": "params_shard_3.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": "8fc38445333e05ba64f309bef4d5f33e" + }, + { + "dataPath": "params_shard_4.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": "63be5165a752874769c14401e8b4de2d" + }, + { + "dataPath": "params_shard_5.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": "dd086ac3ede725c73dfc4d577e5fbf53" + }, + { + "dataPath": "params_shard_6.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": "cdf005d51902a1bd4969ffb337f642ff" + }, + { + "dataPath": "params_shard_7.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": "d4fe2a8fb591d06351fa174c00bde183" + }, + { + "dataPath": "params_shard_8.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": "7515a3ebae036c0a09e5ff72a376a2f9" + }, + { + "dataPath": "params_shard_9.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": "859ff0443caafb7d8e7dbf48414fa594" + }, + { + "dataPath": "params_shard_10.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": "a7f3c447b5da4472b9cdb07a3f37d133" + }, + { + "dataPath": "params_shard_11.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": "0046b7a9dcde345c95dd48ce66f14d68" + }, + { + "dataPath": "params_shard_12.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": "6abc25d45861ffe96b6f0a7422c11cb7" + }, + { + "dataPath": "params_shard_13.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": "f88e90501bd6c51f44524a9c7bee719d" + }, + { + "dataPath": "params_shard_14.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": "0bb8af8147bd7d17f6c001453cd9275b" + }, + { + "dataPath": "params_shard_15.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": "bc9ceba7644f52a8311c3e65a142865d" + }, + { + "dataPath": "params_shard_16.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": "54494bb794b50f6665ad3c4ddca508fb" + }, + { + "dataPath": "params_shard_17.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": "204e4b8d22bb8bc11cefdc99d449362c" + }, + { + "dataPath": "params_shard_18.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": "8dcd8a75799f89ff8adaf4f54e481729" + }, + { + "dataPath": "params_shard_19.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": "5964ec8c6a707776dc19f85d402fa093" + }, + { + "dataPath": "params_shard_20.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": "f00217d78533662404c197fb6018a682" + }, + { + "dataPath": "params_shard_21.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": "41731658ee7d76f3c96b229cc4dcf970" + }, + { + "dataPath": "params_shard_22.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": "d20f6d601b522294043b8f8e359d90c7" + }, + { + "dataPath": "params_shard_23.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": "d8c6be56d9b604c60363ee1bbf5e814a" + }, + { + "dataPath": "params_shard_24.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": "9d5f4e820c8c8b6ebf3b6addf5ecef08" + }, + { + "dataPath": "params_shard_25.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": "fdaa912045247c1c0c4594b9b4447f66" + }, + { + "dataPath": "params_shard_26.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": "44fa09b689ebaa9fba5fdacbd23ba5f1" + }, + { + "dataPath": "params_shard_27.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": "33cd5fdfb18be81cc2fc5ae1c1a00477" + }, + { + "dataPath": "params_shard_28.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": "29dfd7637277c5bc56824415d47603ca" + }, + { + "dataPath": "params_shard_29.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": "39b80c5b31d6c22093c68d2d13400929" + } + ] +} \ No newline at end of file