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": 3.584258148161487 + }, + "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": "c3c69baad78141da371209b3b384fcfd" + }, + { + "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": "894974960779a2dae3d31664d5f0ba27" + }, + { + "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": "29a8219da6436fd89f73ac974a9f03b9" + }, + { + "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": "9c21f07c09bbde969c51cc54c6cee0e2" + }, + { + "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": "ef682c936cc1e24df540a7cf89a060e5" + }, + { + "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": "e9d6e5ee5eff4f9cdbca198a6de5484b" + }, + { + "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": "dcb36737c633544f475af7c570ec9514" + }, + { + "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": "1feecf8066631983863cf9baa368cc9f" + }, + { + "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": "093fa0f7057662306f099f4dfd727961" + }, + { + "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": "2f3c9a5bc0d2aaa197f9a431209b8994" + }, + { + "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": "65d572a8888508d2e04596c57d4e4327" + }, + { + "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": "7824b4ba1682d1952ce856688621dfb2" + }, + { + "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": "1fac41e058793210333c179c5659032e" + }, + { + "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": "5bd4376a7ad1f67da98026414326d0b0" + }, + { + "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": "61e4198da541ef606d50fbc2a8b1a8f4" + }, + { + "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": "36489600441838810fc12ef3b03da733" + }, + { + "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": "2b564cce108f9ce63339e0539d33d1a9" + }, + { + "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": "cca52f13ae86fbd845f5be7f7c8db109" + }, + { + "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": "088e603bea8f8c93d6a752292dd1295f" + }, + { + "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": "40a4d872d43535ba08845205a8e1792e" + }, + { + "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": "30f30b42fab04e15c5816ec485275376" + }, + { + "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": "a86f33fbf01a646c01043c20d706ae74" + }, + { + "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": "da03d8bdbc62364bec8569c837e4a200" + }, + { + "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": "d3a725e6c5defdf169f1dfdcffadf2f0" + }, + { + "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": "452cf880d36a5682a4e7cc110ef77be8" + }, + { + "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": "da99f0482f953a23cfa862b06387d944" + }, + { + "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": "b345281241bef8417cdee734d0ce0994" + }, + { + "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": "d3c40d8f33a12b81dc061c458525b2bf" + }, + { + "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": "e749c9715201814e14c0cdd4cdf3416c" + } + ] +} \ No newline at end of file