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": "16992022e7f3e17597c12a50aadc9738" + }, + { + "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": "28f676f494435cf551ba34147eae4f3e" + }, + { + "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": "0851faf8b47c74d0ff823ba1339e1423" + }, + { + "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": "c2b49e008ac9661dbbdc5c3aa2719325" + }, + { + "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": "e545efac3ce4db0f33069967bb61a68c" + }, + { + "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": "ca7625404e9447497e95900a9a9ba389" + }, + { + "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": "71c29be92595c1d7c75ce6cb5edb757d" + }, + { + "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": "20057fede906dce47877b1dcb40e3799" + }, + { + "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": "43a76d88217f0671e822ea2a611e0910" + }, + { + "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": "7c46a0734b47e048ad463a8b21c09733" + }, + { + "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": "314e510dfe3358f1594557e4a8c02aab" + }, + { + "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": "b0a5537818b8bd56df4e0672bdfc6455" + }, + { + "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": "3989e350e9d994cfce442bf3fc25f2fe" + }, + { + "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": "059d1baf45f875133990d80acb3482a5" + }, + { + "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": "8f14b0af5ab9bcff8a5068acc7823b3d" + }, + { + "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": "3dd2a5b86cff423b64b6edbd64a74aba" + }, + { + "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": "65dfb6ed69b857c4965e5a95673c29ff" + }, + { + "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": "5614217c0713eb645861cd248d80b2f0" + }, + { + "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": "4082969c3158cf5e2ee63f4a0a19e47d" + }, + { + "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": "eecdeffa979decbe6365d62cde7de398" + }, + { + "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": "379bc085e615ac56ed7f5246b9a342c2" + }, + { + "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": "846a9ef6a5e9ab44d15d9c988da643b1" + }, + { + "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": "f3a05da1c09177e0f36b4f4bd06e6d8b" + }, + { + "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": "8fc21cc652c2377afd24e24fcb83e02b" + }, + { + "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": "5ea0b3b3df6bb8507087587beaf75895" + }, + { + "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": "0a773dba0fe99451d2def1bf986b88e5" + }, + { + "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": "63a0adbfc772a74f0f758736d53719c7" + }, + { + "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": "42f44405a35c5e59aca61ed4b438c634" + }, + { + "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": "f8d977a3112146a65a7b45677591d7dd" + } + ] +} \ No newline at end of file