| { | |
| "metadata": { | |
| "ParamSize": 205, | |
| "ParamBytes": 500957200.0, | |
| "BitsPerParam": 32.0 | |
| }, | |
| "records": [ | |
| { | |
| "dataPath": "params_shard_0.bin", | |
| "format": "raw-shard", | |
| "nbytes": 77207040, | |
| "records": [ | |
| { | |
| "name": "roberta.embeddings.word_embeddings.weight", | |
| "shape": [ | |
| 50265, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 77207040, | |
| "byteOffset": 0 | |
| } | |
| ], | |
| "md5sum": "822cb1e659d209d0c4ab502aa588a4a9" | |
| }, | |
| { | |
| "dataPath": "params_shard_1.bin", | |
| "format": "raw-shard", | |
| "nbytes": 32699912, | |
| "records": [ | |
| { | |
| "name": "latency_classifier.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "latency_classifier.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 1536 | |
| }, | |
| { | |
| "name": "latency_classifier.out_proj.bias", | |
| "shape": [ | |
| 2 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4, | |
| "byteOffset": 1181184 | |
| }, | |
| { | |
| "name": "latency_classifier.out_proj.weight", | |
| "shape": [ | |
| 2, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3072, | |
| "byteOffset": 1181188 | |
| }, | |
| { | |
| "name": "quality_classifier.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 1184260 | |
| }, | |
| { | |
| "name": "quality_classifier.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 1185796 | |
| }, | |
| { | |
| "name": "quality_classifier.out_proj.bias", | |
| "shape": [ | |
| 2 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4, | |
| "byteOffset": 2365444 | |
| }, | |
| { | |
| "name": "quality_classifier.out_proj.weight", | |
| "shape": [ | |
| 2, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 3072, | |
| "byteOffset": 2365448 | |
| }, | |
| { | |
| "name": "roberta.embeddings.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 2368520 | |
| }, | |
| { | |
| "name": "roberta.embeddings.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 2370056 | |
| }, | |
| { | |
| "name": "roberta.embeddings.position_embeddings.weight", | |
| "shape": [ | |
| 514, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 789504, | |
| "byteOffset": 2371592 | |
| }, | |
| { | |
| "name": "roberta.embeddings.token_type_embeddings.weight", | |
| "shape": [ | |
| 1, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 3161096 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 3162632 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 3164168 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 3165704 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 3167240 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4346888 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 4348424 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 5528072 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 5529608 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 6709256 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 6710792 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 7890440 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 7896584 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 12615176 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 12616712 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 12618248 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.0.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 12619784 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 17338376 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 17339912 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 17341448 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 17342984 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 18522632 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 18524168 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 19703816 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 19705352 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 20885000 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 20886536 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 22066184 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 22072328 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 26790920 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 26792456 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 26793992 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.1.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 26795528 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 31514120 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 31515656 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 31517192 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 31518728 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 32698376 | |
| } | |
| ], | |
| "md5sum": "c773b3066463e8b7ca2c687935fdd38d" | |
| }, | |
| { | |
| "dataPath": "params_shard_2.bin", | |
| "format": "raw-shard", | |
| "nbytes": 31899648, | |
| "records": [ | |
| { | |
| "name": "roberta.encoder.layer.10.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 1179648 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 1181184 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 2360832 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 2362368 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 3542016 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 3548160 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 8266752 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 8268288 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 8269824 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.10.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 8271360 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 12989952 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 12991488 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 12993024 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 12994560 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 14174208 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 14175744 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 15355392 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 15356928 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 16536576 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 16538112 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 17717760 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 17723904 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 22442496 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 22444032 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 22445568 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.11.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 22447104 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 27165696 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 27167232 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 27168768 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 27170304 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 28349952 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 28351488 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 29531136 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 29532672 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 30712320 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 30713856 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 31893504 | |
| } | |
| ], | |
| "md5sum": "cd967b7aa914de66a898f89fcac8f9dc" | |
| }, | |
| { | |
| "dataPath": "params_shard_3.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33074688, | |
| "records": [ | |
| { | |
| "name": "roberta.encoder.layer.2.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4718592 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4720128 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4721664 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.2.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 4723200 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 9441792 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 9443328 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 9444864 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 9446400 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 10626048 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 10627584 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 11807232 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 11808768 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 12988416 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 12989952 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 14169600 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 14175744 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 18894336 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 18895872 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 18897408 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.3.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 18898944 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 23617536 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 23619072 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 23620608 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 23622144 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 24801792 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 24803328 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 25982976 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 25984512 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 27164160 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 27165696 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 28345344 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 28351488 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 33070080 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 33071616 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.4.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 33073152 | |
| } | |
| ], | |
| "md5sum": "f1598d0a023d40021c0400ee9f49844d" | |
| }, | |
| { | |
| "dataPath": "params_shard_4.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33074688, | |
| "records": [ | |
| { | |
| "name": "roberta.encoder.layer.4.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4718592 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4720128 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4721664 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 4723200 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 5902848 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 5904384 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 7084032 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 7085568 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 8265216 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 8266752 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 9446400 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 9452544 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 14171136 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 14172672 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 14174208 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.5.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 14175744 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 18894336 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 18895872 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 18897408 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 18898944 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 20078592 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 20080128 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 21259776 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 21261312 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 22440960 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 22442496 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 23622144 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 23628288 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 28346880 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 28348416 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 28349952 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.6.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 28351488 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 33070080 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 33071616 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 33073152 | |
| } | |
| ], | |
| "md5sum": "d7db2ff4e87fb3b13f5a0f5e9c00d5e9" | |
| }, | |
| { | |
| "dataPath": "params_shard_5.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33080832, | |
| "records": [ | |
| { | |
| "name": "roberta.encoder.layer.7.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 1179648 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 1181184 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 2360832 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 2362368 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 3542016 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 3543552 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 4723200 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 4729344 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 9447936 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 9449472 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 9451008 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.7.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 9452544 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 14171136 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 14172672 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 14174208 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 14175744 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 15355392 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 15356928 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 16536576 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 16538112 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 17717760 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 17719296 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 18898944 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 18905088 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 23623680 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 23625216 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 23626752 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.8.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 23628288 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 28346880 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 28348416 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 28349952 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 28351488 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.self.key.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 29531136 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.self.key.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 29532672 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.self.query.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 30712320 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.self.query.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 30713856 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.self.value.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 31893504 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.attention.self.value.weight", | |
| "shape": [ | |
| 768, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1179648, | |
| "byteOffset": 31895040 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.intermediate.dense.bias", | |
| "shape": [ | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 33074688 | |
| } | |
| ], | |
| "md5sum": "16e5fe6cb9c4b8dfe68170b6b4a350be" | |
| }, | |
| { | |
| "dataPath": "params_shard_6.bin", | |
| "format": "raw-shard", | |
| "nbytes": 9441792, | |
| "records": [ | |
| { | |
| "name": "roberta.encoder.layer.9.intermediate.dense.weight", | |
| "shape": [ | |
| 3072, | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.output.LayerNorm.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4718592 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.output.LayerNorm.weight", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4720128 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.output.dense.bias", | |
| "shape": [ | |
| 768 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1536, | |
| "byteOffset": 4721664 | |
| }, | |
| { | |
| "name": "roberta.encoder.layer.9.output.dense.weight", | |
| "shape": [ | |
| 768, | |
| 3072 | |
| ], | |
| "dtype": "float32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 4718592, | |
| "byteOffset": 4723200 | |
| } | |
| ], | |
| "md5sum": "5882c18b7fb291d522cc147ec14e7014" | |
| } | |
| ] | |
| } |