| { | |
| "metadata": { | |
| "ParamSize": 455, | |
| "ParamBytes": 1609550312.0, | |
| "BitsPerParam": 4.50391353017126 | |
| }, | |
| "records": [ | |
| { | |
| "dataPath": "params_shard_0.bin", | |
| "format": "raw-shard", | |
| "nbytes": 85345280, | |
| "records": [ | |
| { | |
| "name": "lm_head.linear.q_weight", | |
| "shape": [ | |
| 66676, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 85345280, | |
| "byteOffset": 0 | |
| } | |
| ], | |
| "md5sum": "6a6a374812c778dd82b56bb4c3a717cb" | |
| }, | |
| { | |
| "dataPath": "params_shard_1.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25593192, | |
| "records": [ | |
| { | |
| "name": "lm_head.linear.bias", | |
| "shape": [ | |
| 66676 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 133352, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "lm_head.linear.q_scale", | |
| "shape": [ | |
| 66676, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 10668160, | |
| "byteOffset": 133352 | |
| }, | |
| { | |
| "name": "lm_head.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 10801512 | |
| }, | |
| { | |
| "name": "lm_head.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 10806632 | |
| }, | |
| { | |
| "name": "transformer.h.29.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 10811752 | |
| }, | |
| { | |
| "name": "transformer.h.29.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 10816872 | |
| }, | |
| { | |
| "name": "transformer.h.29.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 10821992 | |
| }, | |
| { | |
| "name": "transformer.h.29.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 10842472 | |
| }, | |
| { | |
| "name": "transformer.h.29.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 23949672 | |
| }, | |
| { | |
| "name": "transformer.h.29.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25588072 | |
| } | |
| ], | |
| "md5sum": "b466ad7ae54e2fd70a8bf3bf9e942373" | |
| }, | |
| { | |
| "dataPath": "params_shard_2.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29527040, | |
| "records": [ | |
| { | |
| "name": "transformer.h.29.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.29.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.30.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.30.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.30.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 14755840 | |
| }, | |
| { | |
| "name": "transformer.h.30.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14776320 | |
| }, | |
| { | |
| "name": "transformer.h.30.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27883520 | |
| }, | |
| { | |
| "name": "transformer.h.30.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "e720dfcfed421199e581c5bb476e1705" | |
| }, | |
| { | |
| "dataPath": "params_shard_3.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.30.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.30.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.30.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.30.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.30.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.30.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.30.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.30.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.31.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.31.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.31.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "864c61b23e4a7ac68de0ac5eb2fb09c4" | |
| }, | |
| { | |
| "dataPath": "params_shard_4.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.31.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.31.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.31.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.31.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.31.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.31.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.31.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.31.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.31.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "3093e47a9d3a93e759428e38e285a7af" | |
| }, | |
| { | |
| "dataPath": "params_shard_5.bin", | |
| "format": "raw-shard", | |
| "nbytes": 85345280, | |
| "records": [ | |
| { | |
| "name": "transformer.embd.q_weight", | |
| "shape": [ | |
| 66676, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 85345280, | |
| "byteOffset": 0 | |
| } | |
| ], | |
| "md5sum": "e904e988708568193326c3b863d24e7b" | |
| }, | |
| { | |
| "dataPath": "params_shard_6.bin", | |
| "format": "raw-shard", | |
| "nbytes": 21758080, | |
| "records": [ | |
| { | |
| "name": "transformer.h.31.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.31.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.embd.q_scale", | |
| "shape": [ | |
| 66676, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 10668160, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.0.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 21727360 | |
| }, | |
| { | |
| "name": "transformer.h.0.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 21732480 | |
| }, | |
| { | |
| "name": "transformer.h.0.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 21737600 | |
| } | |
| ], | |
| "md5sum": "92da578d62a95582669d859fa57ee360" | |
| }, | |
| { | |
| "dataPath": "params_shard_7.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.0.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.0.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.0.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.0.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.0.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.0.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.0.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.0.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.0.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "ae891b8d27ad9fa41c233b568d8c47d3" | |
| }, | |
| { | |
| "dataPath": "params_shard_8.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.0.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.0.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.1.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.1.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.1.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.1.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.1.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.1.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "a164e53821c5f92a9e3fdfa655ae3fbc" | |
| }, | |
| { | |
| "dataPath": "params_shard_9.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.1.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.1.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.1.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.1.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.1.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.1.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.1.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.1.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.10.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.10.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.10.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "60b7ffd06a562b5998625c763cd7d86d" | |
| }, | |
| { | |
| "dataPath": "params_shard_10.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.10.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.10.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.10.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.10.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.10.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.10.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.10.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.10.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.10.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "5c4b89399210a6500581d4b27adbabe9" | |
| }, | |
| { | |
| "dataPath": "params_shard_11.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.10.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.10.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.11.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.11.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.11.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.11.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.11.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.11.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "4b15c4874786e52c0d9755b6073eb76e" | |
| }, | |
| { | |
| "dataPath": "params_shard_12.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.11.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.11.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.11.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.11.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.11.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.11.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.11.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.11.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.12.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.12.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.12.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "61d0be0db3fa697989f37009c099aecd" | |
| }, | |
| { | |
| "dataPath": "params_shard_13.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.12.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.12.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.12.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.12.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.12.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.12.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.12.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.12.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.12.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "a8a776082f811f27f6ea328381745d15" | |
| }, | |
| { | |
| "dataPath": "params_shard_14.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.12.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.12.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.13.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.13.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.13.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.13.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.13.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.13.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "1af7ba9771871e411a4ed8d4347e4b14" | |
| }, | |
| { | |
| "dataPath": "params_shard_15.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.13.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.13.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.13.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.13.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.13.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.13.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.13.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.13.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.14.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.14.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.14.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "211c7fd26b4312c920606353eb66e975" | |
| }, | |
| { | |
| "dataPath": "params_shard_16.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.14.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.14.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.14.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.14.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.14.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.14.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.14.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.14.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.14.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "21e497b6126fe84a1fe0eb9f8b564ff8" | |
| }, | |
| { | |
| "dataPath": "params_shard_17.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.14.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.14.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.15.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.15.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.15.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.15.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.15.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.15.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "fb57c874f1770e3a5cbb419cb65cadc6" | |
| }, | |
| { | |
| "dataPath": "params_shard_18.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.15.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.15.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.15.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.15.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.15.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.15.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.15.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.15.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.16.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.16.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.16.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "650a61e3b031144b5b7277c847775261" | |
| }, | |
| { | |
| "dataPath": "params_shard_19.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.16.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.16.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.16.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.16.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.16.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.16.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.16.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.16.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.16.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "e170f65631a3851d027df4111cfd1689" | |
| }, | |
| { | |
| "dataPath": "params_shard_20.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.16.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.16.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.17.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.17.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.17.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.17.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.17.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.17.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "9acf0373a1e16c933fe80be59509f369" | |
| }, | |
| { | |
| "dataPath": "params_shard_21.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.17.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.17.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.17.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.17.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.17.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.17.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.17.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.17.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.18.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.18.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.18.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "92b35718640d7746279f6d93b0b146c8" | |
| }, | |
| { | |
| "dataPath": "params_shard_22.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.18.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.18.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.18.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.18.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.18.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.18.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.18.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.18.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.18.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "df8fd6127484d74df52cda24aa0aafab" | |
| }, | |
| { | |
| "dataPath": "params_shard_23.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.18.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.18.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.19.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.19.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.19.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.19.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.19.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.19.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "35a77b8ae388147a1826a957766efde8" | |
| }, | |
| { | |
| "dataPath": "params_shard_24.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.19.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.19.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.19.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.19.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.19.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.19.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.19.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.19.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.2.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.2.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.2.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "294b21a0989e6d9c6ca37c2fc0017d3e" | |
| }, | |
| { | |
| "dataPath": "params_shard_25.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.2.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.2.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.2.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.2.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.2.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.2.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.2.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.2.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.2.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "9bbfe1492b2fd3a888b0b51731a555e2" | |
| }, | |
| { | |
| "dataPath": "params_shard_26.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.2.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.2.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.20.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.20.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.20.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.20.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.20.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.20.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "c58f606dbecf51400073294631a70ff1" | |
| }, | |
| { | |
| "dataPath": "params_shard_27.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.20.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.20.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.20.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.20.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.20.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.20.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.20.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.20.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.21.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.21.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.21.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "66d983a402076475e4774da1b94213e0" | |
| }, | |
| { | |
| "dataPath": "params_shard_28.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.21.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.21.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.21.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.21.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.21.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.21.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.21.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.21.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.21.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "72496b6c67f11b9dfb0cda6e1038ff1e" | |
| }, | |
| { | |
| "dataPath": "params_shard_29.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.21.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.21.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.22.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.22.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.22.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.22.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.22.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.22.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "6a1f8a60760f9aae6e8e30ed8f0a4980" | |
| }, | |
| { | |
| "dataPath": "params_shard_30.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.22.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.22.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.22.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.22.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.22.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.22.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.22.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.22.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.23.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.23.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.23.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "99f80b07fc1effe70991e9b5e6325d55" | |
| }, | |
| { | |
| "dataPath": "params_shard_31.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.23.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.23.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.23.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.23.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.23.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.23.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.23.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.23.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.23.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "dbbff54c9fa8ef4968571134cf28f1ea" | |
| }, | |
| { | |
| "dataPath": "params_shard_32.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.23.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.23.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.24.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.24.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.24.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.24.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.24.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.24.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "d225222ca4865d782ba151c4093935b3" | |
| }, | |
| { | |
| "dataPath": "params_shard_33.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.24.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.24.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.24.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.24.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.24.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.24.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.24.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.24.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.25.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.25.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.25.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "ed4b70384e5ae4b9a65cab6a2ebca953" | |
| }, | |
| { | |
| "dataPath": "params_shard_34.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.25.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.25.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.25.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.25.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.25.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.25.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.25.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.25.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.25.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "70e4ec156bf47504b0d1afce0872eb04" | |
| }, | |
| { | |
| "dataPath": "params_shard_35.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.25.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.25.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.26.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.26.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.26.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.26.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.26.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.26.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "31873947af864d887460a44f36c3dc45" | |
| }, | |
| { | |
| "dataPath": "params_shard_36.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.26.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.26.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.26.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.26.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.26.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.26.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.26.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.26.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.27.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.27.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.27.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "149a8ac1f416d76466299352ced76d99" | |
| }, | |
| { | |
| "dataPath": "params_shard_37.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.27.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.27.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.27.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.27.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.27.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.27.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.27.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.27.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.27.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "5cb411d96e9827234fa9ec5cf6035fbd" | |
| }, | |
| { | |
| "dataPath": "params_shard_38.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.27.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.27.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.28.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.28.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.28.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.28.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.28.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.28.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "c107297bd9e1ce850de7a1edd55f524e" | |
| }, | |
| { | |
| "dataPath": "params_shard_39.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33218560, | |
| "records": [ | |
| { | |
| "name": "transformer.h.28.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.28.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.28.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.28.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.28.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.28.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.28.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.28.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.29.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.29.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.29.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32793600 | |
| }, | |
| { | |
| "name": "transformer.h.29.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33203200 | |
| } | |
| ], | |
| "md5sum": "687bcb58c9e1e032878bb5b64cdd01e3" | |
| }, | |
| { | |
| "dataPath": "params_shard_40.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.29.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.29.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.3.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.3.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.3.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.3.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.3.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.3.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "ccc43d49c7e9ba1906f72615dc8fe07d" | |
| }, | |
| { | |
| "dataPath": "params_shard_41.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.3.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.3.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.3.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.3.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.3.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.3.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.3.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.3.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.4.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.4.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.4.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "6487b64be34dc264dee04225857763db" | |
| }, | |
| { | |
| "dataPath": "params_shard_42.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.4.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.4.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.4.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.4.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.4.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.4.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.4.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.4.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.4.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "07508e14c3899fefe6ee92a7ba762fcc" | |
| }, | |
| { | |
| "dataPath": "params_shard_43.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.4.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.4.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.5.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.5.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.5.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.5.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.5.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.5.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "975bf8fa6a1ab10966021c5b06f822ae" | |
| }, | |
| { | |
| "dataPath": "params_shard_44.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.5.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.5.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.5.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.5.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.5.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.5.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.5.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.5.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.6.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.6.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.6.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "60c4e58d5ff85e49b5323f41ca86f901" | |
| }, | |
| { | |
| "dataPath": "params_shard_45.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.6.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.6.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.6.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.6.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.6.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.6.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.6.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.6.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.6.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "2e50dec2540863088e00077503125c42" | |
| }, | |
| { | |
| "dataPath": "params_shard_46.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.6.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.6.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.7.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.7.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.7.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.7.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.7.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.7.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "b050c360ea7d833cc7eca6fbde0c73c0" | |
| }, | |
| { | |
| "dataPath": "params_shard_47.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29542400, | |
| "records": [ | |
| { | |
| "name": "transformer.h.7.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.7.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.7.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.7.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.7.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.7.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.7.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.7.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| }, | |
| { | |
| "name": "transformer.h.8.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29511680 | |
| }, | |
| { | |
| "name": "transformer.h.8.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29516800 | |
| }, | |
| { | |
| "name": "transformer.h.8.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 29521920 | |
| } | |
| ], | |
| "md5sum": "6d49f126a69a884df1f2fa3b404aedfc" | |
| }, | |
| { | |
| "dataPath": "params_shard_48.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33203200, | |
| "records": [ | |
| { | |
| "name": "transformer.h.8.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.8.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.8.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.8.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.8.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 27857920 | |
| }, | |
| { | |
| "name": "transformer.h.8.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 29496320 | |
| }, | |
| { | |
| "name": "transformer.h.8.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 29501440 | |
| }, | |
| { | |
| "name": "transformer.h.8.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 32778240 | |
| }, | |
| { | |
| "name": "transformer.h.8.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 33187840 | |
| } | |
| ], | |
| "md5sum": "d13cce06be0e9ab8efbd576ecf14da9d" | |
| }, | |
| { | |
| "dataPath": "params_shard_49.bin", | |
| "format": "raw-shard", | |
| "nbytes": 25840640, | |
| "records": [ | |
| { | |
| "name": "transformer.h.8.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.8.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 9830400 | |
| }, | |
| { | |
| "name": "transformer.h.9.ln.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11059200 | |
| }, | |
| { | |
| "name": "transformer.h.9.ln.weight", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 11064320 | |
| }, | |
| { | |
| "name": "transformer.h.9.mlp.fc1.bias", | |
| "shape": [ | |
| 10240 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 20480, | |
| "byteOffset": 11069440 | |
| }, | |
| { | |
| "name": "transformer.h.9.mlp.fc1.q_weight", | |
| "shape": [ | |
| 10240, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 11089920 | |
| }, | |
| { | |
| "name": "transformer.h.9.mlp.fc1.q_scale", | |
| "shape": [ | |
| 10240, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 24197120 | |
| }, | |
| { | |
| "name": "transformer.h.9.mlp.fc2.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 25835520 | |
| } | |
| ], | |
| "md5sum": "85ad4588abe3b60680d971acf88e755b" | |
| }, | |
| { | |
| "dataPath": "params_shard_50.bin", | |
| "format": "raw-shard", | |
| "nbytes": 29511680, | |
| "records": [ | |
| { | |
| "name": "transformer.h.9.mlp.fc2.q_weight", | |
| "shape": [ | |
| 2560, | |
| 1280 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 13107200, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "transformer.h.9.mlp.fc2.q_scale", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1638400, | |
| "byteOffset": 13107200 | |
| }, | |
| { | |
| "name": "transformer.h.9.mixer.out_proj.bias", | |
| "shape": [ | |
| 2560 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 5120, | |
| "byteOffset": 14745600 | |
| }, | |
| { | |
| "name": "transformer.h.9.mixer.out_proj.q_weight", | |
| "shape": [ | |
| 2560, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3276800, | |
| "byteOffset": 14750720 | |
| }, | |
| { | |
| "name": "transformer.h.9.mixer.out_proj.q_scale", | |
| "shape": [ | |
| 2560, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 409600, | |
| "byteOffset": 18027520 | |
| }, | |
| { | |
| "name": "transformer.h.9.mixer.Wqkv.bias", | |
| "shape": [ | |
| 7680 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 15360, | |
| "byteOffset": 18437120 | |
| }, | |
| { | |
| "name": "transformer.h.9.mixer.Wqkv.q_weight", | |
| "shape": [ | |
| 7680, | |
| 320 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 9830400, | |
| "byteOffset": 18452480 | |
| }, | |
| { | |
| "name": "transformer.h.9.mixer.Wqkv.q_scale", | |
| "shape": [ | |
| 7680, | |
| 80 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1228800, | |
| "byteOffset": 28282880 | |
| } | |
| ], | |
| "md5sum": "807ef5af755d46b2aea6601de08c8d5f" | |
| } | |
| ] | |
| } |