| { | |
| "prepared_at_utc": "2026-10-26T05:53:07.208785+00:00", | |
| "encoder": "Exact bundled GPT-2/r50k via wiggle_assistant.get_encoder; ordinary-string encoding only", | |
| "tokenizer_source": "original_workspace/WiggleGPT-Gosport/wiggle_assistant.py", | |
| "tokenizer_source_sha256": "fc10c18e3cd43986437c0f2ecde5a6b1c167423d9db98243e6f8c7e393711791", | |
| "merge_table_sha256": "306cd27f03c1a714eca7108e03d66b7dc042abe8c258b44c199a7ed9838dd930", | |
| "task_markers": { | |
| "START": 50257, | |
| "DELIM": 50258, | |
| "CLS": 50259, | |
| "PAD": 50256 | |
| }, | |
| "limit": 512, | |
| "cpu_threads": 2, | |
| "splits": { | |
| "sst2_train": { | |
| "task": "sst2", | |
| "split": "train", | |
| "examples": 67349, | |
| "choices": 1, | |
| "source_jsonl": "original_workspace/output/gpt1_style/task_finetuning/data/sst2_train.jsonl", | |
| "source_sha256": "75fd14e7a62a5d54d1300996afaef6223ec8c49eecd97f0674d6664489a69daa", | |
| "pad_length": 80, | |
| "max_sequence_length": 67, | |
| "min_sequence_length": 4, | |
| "max_untruncated_length": 67, | |
| "truncation": { | |
| "examples_with_truncation": 0, | |
| "candidate_sequences_with_truncation": 0, | |
| "total_tokens_removed": 0, | |
| "max_tokens_removed_per_example": 0 | |
| }, | |
| "context_policy": "[START] + review first 510 tokens + [CLS]; right-pad to the split maximum rounded up to a multiple of 16 (minimum 16).", | |
| "input_ids_dtype": "int32", | |
| "lengths_dtype": "int64", | |
| "labels_dtype": "int64", | |
| "cache_path": "original_workspace/output/gpt1_style/task_finetuning/data/sst2_train.pt", | |
| "cache_sha256": "eacd114353b3cd0a0a89490a2d81e54f4ddc95d65cd74efb143b50c7b5581e57", | |
| "cache_bytes": 24371735 | |
| }, | |
| "sst2_validation": { | |
| "task": "sst2", | |
| "split": "validation", | |
| "examples": 872, | |
| "choices": 1, | |
| "source_jsonl": "original_workspace/output/gpt1_style/task_finetuning/data/sst2_validation.jsonl", | |
| "source_sha256": "9f7afa11e8d0309f4ea4732c7f11d78fb0f686f48fac5596a06db2ecf496678b", | |
| "pad_length": 64, | |
| "max_sequence_length": 63, | |
| "min_sequence_length": 5, | |
| "max_untruncated_length": 63, | |
| "truncation": { | |
| "examples_with_truncation": 0, | |
| "candidate_sequences_with_truncation": 0, | |
| "total_tokens_removed": 0, | |
| "max_tokens_removed_per_example": 0 | |
| }, | |
| "context_policy": "[START] + review first 510 tokens + [CLS]; right-pad to the split maximum rounded up to a multiple of 16 (minimum 16).", | |
| "input_ids_dtype": "int32", | |
| "lengths_dtype": "int64", | |
| "labels_dtype": "int64", | |
| "cache_path": "original_workspace/output/gpt1_style/task_finetuning/data/sst2_validation.pt", | |
| "cache_sha256": "22eed8372c39df075210f60de11263bfd6f8dd17cc934901e0d1f807410bf943", | |
| "cache_bytes": 264900 | |
| }, | |
| "sst2_test": { | |
| "task": "sst2", | |
| "split": "test", | |
| "examples": 1821, | |
| "choices": 1, | |
| "source_jsonl": "original_workspace/output/gpt1_style/task_finetuning/data/sst2_test.jsonl", | |
| "source_sha256": "2b8c1efc9ebea4f988382cc5d91aa3b636bff90e16795d705b2a25ef8b2cfd5b", | |
| "pad_length": 64, | |
| "max_sequence_length": 63, | |
| "min_sequence_length": 5, | |
| "max_untruncated_length": 63, | |
| "truncation": { | |
| "examples_with_truncation": 0, | |
| "candidate_sequences_with_truncation": 0, | |
| "total_tokens_removed": 0, | |
| "max_tokens_removed_per_example": 0 | |
| }, | |
| "context_policy": "[START] + review first 510 tokens + [CLS]; right-pad to the split maximum rounded up to a multiple of 16 (minimum 16).", | |
| "input_ids_dtype": "int32", | |
| "lengths_dtype": "int64", | |
| "labels_dtype": "int64", | |
| "cache_path": "original_workspace/output/gpt1_style/task_finetuning/data/sst2_test.pt", | |
| "cache_sha256": "577a4c443c0bf54dd01c12d63c0b5a485d455256d741a740363791db6298cdd0", | |
| "cache_bytes": 540302 | |
| }, | |
| "race_middle_train": { | |
| "task": "race_middle", | |
| "split": "train", | |
| "examples": 25421, | |
| "choices": 4, | |
| "source_jsonl": "original_workspace/output/gpt1_style/task_finetuning/data/race_middle_train.jsonl", | |
| "source_sha256": "ed805e44773ff5856211f0713270222c8ca681e84c43b549cdecfbef97a46652", | |
| "pad_length": 512, | |
| "max_sequence_length": 512, | |
| "min_sequence_length": 20, | |
| "max_untruncated_length": 673, | |
| "truncation": { | |
| "examples_with_truncation": 210, | |
| "candidate_sequences_with_truncation": 840, | |
| "total_tokens_removed": 10052, | |
| "max_tokens_removed_per_example": 161 | |
| }, | |
| "context_policy": "[START] + shared article prefix + [DELIM] + complete question + [DELIM] + complete answer + [CLS]; all four choices retain the same article prefix, budgeted using the longest answer; right-pad to 512.", | |
| "input_ids_dtype": "int32", | |
| "lengths_dtype": "int64", | |
| "labels_dtype": "int64", | |
| "cache_path": "original_workspace/output/gpt1_style/task_finetuning/data/race_middle_train.pt", | |
| "cache_sha256": "5494c6d1a4fcca3651e819680e094693db032a36991d4b1341b108d7d0c5a797", | |
| "cache_bytes": 209874966 | |
| }, | |
| "race_middle_validation": { | |
| "task": "race_middle", | |
| "split": "validation", | |
| "examples": 1436, | |
| "choices": 4, | |
| "source_jsonl": "original_workspace/output/gpt1_style/task_finetuning/data/race_middle_validation.jsonl", | |
| "source_sha256": "385d417aaddbc25038d7bd4afbf9f385d4e3ae74bff532397ba938b0fbf0f622", | |
| "pad_length": 512, | |
| "max_sequence_length": 512, | |
| "min_sequence_length": 65, | |
| "max_untruncated_length": 530, | |
| "truncation": { | |
| "examples_with_truncation": 9, | |
| "candidate_sequences_with_truncation": 36, | |
| "total_tokens_removed": 71, | |
| "max_tokens_removed_per_example": 18 | |
| }, | |
| "context_policy": "[START] + shared article prefix + [DELIM] + complete question + [DELIM] + complete answer + [CLS]; all four choices retain the same article prefix, budgeted using the longest answer; right-pad to 512.", | |
| "input_ids_dtype": "int32", | |
| "lengths_dtype": "int64", | |
| "labels_dtype": "int64", | |
| "cache_path": "original_workspace/output/gpt1_style/task_finetuning/data/race_middle_validation.pt", | |
| "cache_sha256": "fb2db9a134abcf138b093d251172c6ffde4e58db60a0c2a56657956ed9afeae6", | |
| "cache_bytes": 11858115 | |
| }, | |
| "race_middle_test": { | |
| "task": "race_middle", | |
| "split": "test", | |
| "examples": 1436, | |
| "choices": 4, | |
| "source_jsonl": "original_workspace/output/gpt1_style/task_finetuning/data/race_middle_test.jsonl", | |
| "source_sha256": "dd9756cf41d600260b59064c2b10395affb53d769a4bf71017ca8dc6ceb005d9", | |
| "pad_length": 512, | |
| "max_sequence_length": 512, | |
| "min_sequence_length": 41, | |
| "max_untruncated_length": 565, | |
| "truncation": { | |
| "examples_with_truncation": 14, | |
| "candidate_sequences_with_truncation": 56, | |
| "total_tokens_removed": 531, | |
| "max_tokens_removed_per_example": 53 | |
| }, | |
| "context_policy": "[START] + shared article prefix + [DELIM] + complete question + [DELIM] + complete answer + [CLS]; all four choices retain the same article prefix, budgeted using the longest answer; right-pad to 512.", | |
| "input_ids_dtype": "int32", | |
| "lengths_dtype": "int64", | |
| "labels_dtype": "int64", | |
| "cache_path": "original_workspace/output/gpt1_style/task_finetuning/data/race_middle_test.pt", | |
| "cache_sha256": "7d5577758ee2fbd2429edc03669d2f539568158d6658683c8f872ce400375e56", | |
| "cache_bytes": 11857997 | |
| } | |
| } | |
| } | |