| # Opsis V1 snapshot fingerprint |
|
|
| The record below identifies the local files used to train and evaluate the published Opsis checkpoint. The restricted training images and row manifests are not redistributed here. |
|
|
| ```json |
| { |
| "schema_version": 1, |
| "scope": "Public provenance record only. No source images, generated images, annotations, training outputs, or row-level manifests are redistributed.", |
| "model_repo": "yafitzdev/opsis-v1-nano-g3", |
| "model_revision_at_packaging": "99f24dadea6f1d260be190d6b34a7f47fce8318a", |
| "model_safetensors_sha256": "c05926f665413883b0310a571f23efdd546ceda6e5e6fd06655f5af3173310a1", |
| "training_rows": 100000, |
| "validation_rows": 1000, |
| "training_unique_image_paths": 96875, |
| "training_repeated_image_rows": 3125, |
| "image_set_including_validation": { |
| "algorithm": "sha256 over sorted relative POSIX path, NUL, decimal byte length, NUL, file bytes for each unique image", |
| "unique_image_files": 97875, |
| "total_image_bytes": 20691179759, |
| "digest": "5c90edf64db06d70eaff06cb5ea6c193b57aa8f6e3f13cb76f67fec86dbf8dcf", |
| "rows": { |
| "train": 100000, |
| "validation": 1000 |
| } |
| }, |
| "training_content_counts": { |
| "table": 31250, |
| "chart": 18750, |
| "diagram": 18750, |
| "ordinary_image": 31250 |
| }, |
| "training_domain_counts": { |
| "clean": 75000, |
| "pdf_style_crop": 25000 |
| }, |
| "gold_benchmark_cases": 600, |
| "gold_training_overlap_reported_by_curation": 0, |
| "manifests": { |
| "train_manifest": { |
| "local_path": "artifacts/public_data/training_g3_100k_corrected/train.jsonl", |
| "bytes": 54468933, |
| "sha256": "f86fbf08a9583be98a4b6a29a72d9146af6cab29eb40043cc236615625b96e90" |
| }, |
| "validation_manifest": { |
| "local_path": "artifacts/public_data/training_g3_100k_corrected/validation.jsonl", |
| "bytes": 497277, |
| "sha256": "1890ad974ca611836ae93c6a6e3be87a86089b63997fe9d508135da81b919a44" |
| }, |
| "corrected_report": { |
| "local_path": "artifacts/public_data/training_g3_100k_corrected/report.json", |
| "bytes": 592, |
| "sha256": "74ff45c2ed042b83d11a1c0c3e6e8a9e3a1a65f6cc7e6651d218077f4a1f32f1" |
| }, |
| "source_curation_report": { |
| "local_path": "artifacts/public_data/training_g3_100k/curation_report.json", |
| "bytes": 4980, |
| "sha256": "02a9962d7a52e468b66553ec7a63950e9e349c31d7b513faf3852516c21fff4f" |
| }, |
| "source_download_manifest": { |
| "local_path": "artifacts/public_data/raw/download_manifest.json", |
| "bytes": 2918, |
| "sha256": "e78370dd33d7f32ad7193523480eb7b3f2b98864621c557e4a34435139b97b63" |
| }, |
| "curator_code": { |
| "local_path": "scripts/curate_training_g3.py", |
| "bytes": 15038, |
| "sha256": "9e7f1afb2478df6360d1dba4aeb8c31aa07acff20fe3803dca05a183a63c158b" |
| }, |
| "correction_code": { |
| "local_path": "scripts/make_g3_corrected_manifest.py", |
| "bytes": 5615, |
| "sha256": "fc7db47e8ede4b966582e2c06e4b97de072cd8c5ed17e3e9153b12736bdedcec" |
| } |
| }, |
| "source_archives": [ |
| { |
| "id": "ai2d", |
| "source_url": "https://ai2-website.s3.amazonaws.com/data/ai2d-all.zip", |
| "upstream": "https://prior.allenai.org/projects/diagram-understanding", |
| "archive_bytes": 990965374, |
| "archive_sha256": "1a6b77eebb8b7dbdf76a0ba6ca76c2f97ce8f81d8ee33b06593aa722e54c4786", |
| "license_recorded_by_project": "AI2D research-only license (non-commercial; no redistribution)" |
| }, |
| { |
| "id": "chartqa", |
| "source_url": "https://huggingface.co/datasets/ahmed-masry/ChartQA/resolve/af8b6f5c08c95085271561c2a3f9d15f2b5a9031/ChartQA%20Dataset.zip?download=true", |
| "upstream": "https://github.com/vis-nlp/ChartQA", |
| "archive_bytes": 875370872, |
| "archive_sha256": "1bf310e5a51101681495c4a24f4f29d22c4f70b52df24e2e4feb0d79cae3c160", |
| "license_recorded_by_project": "GPL-3.0" |
| }, |
| { |
| "id": "docci_annotations", |
| "source_url": "https://storage.googleapis.com/docci/data/docci_descriptions.jsonlines", |
| "upstream": "https://google.github.io/docci/", |
| "archive_bytes": 11000214, |
| "archive_sha256": "c9df4819963883af35ddd2cf257949892fd8c6d88b33a012094352df60719800", |
| "license_recorded_by_project": "CC-BY-4.0" |
| }, |
| { |
| "id": "docci_images", |
| "source_url": "https://storage.googleapis.com/docci/data/docci_images.tar.gz", |
| "upstream": "https://google.github.io/docci/", |
| "archive_bytes": 7592938768, |
| "archive_sha256": "c1b1aee00856757d71cfe2cc6ab641089284276a791bfb82146b4742c59a0a1d", |
| "license_recorded_by_project": "CC-BY-4.0" |
| }, |
| { |
| "id": "pubtabnet", |
| "source_url": "https://huggingface.co/datasets/ajimeno/PubTabNet/resolve/06963b1af16203f4633c718e2c50109eb57ed658/pubtabnet.tar.gz?download=true", |
| "upstream": "https://github.com/ibm-aur-nlp/PubTabNet", |
| "archive_bytes": 11244059914, |
| "archive_sha256": "90c55e733c85c98edf6d350b77f1e4c23767577555fc932e44ad723674de8d3e", |
| "license_recorded_by_project": "CDLA-Sharing-1.0" |
| } |
| ], |
| "redistribution_boundary": "The bundled AI2D license bars distribution of its data to third parties and public release of modified data. Full training examples are deliberately absent." |
| } |
| ``` |
|
|
