| { |
| "schemaVersion": 1, |
| "updatedAt": "2026-07-11", |
| "models": [ |
| { |
| "modelId": "ppocr-v6-tiny-det", |
| "role": "ocr.det", |
| "family": "PP-OCRv6_tiny_det", |
| "upstreamModelFamily": "PP-OCRv6_tiny_det", |
| "source": { |
| "kind": "huggingface-paddle-static", |
| "repo": "PaddlePaddle/PP-OCRv6_tiny_det", |
| "revision": "d3177d4e5551463292a61e27cfca2b53e7c3fe9d", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "inference.json", |
| "sizeBytes": 229184, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "56ddf4c08ad72e179d03a0caffdf794df5ffa47c" |
| }, |
| "purpose": "Paddle static graph program description" |
| }, |
| { |
| "path": "inference.pdiparams", |
| "sizeBytes": 1718055, |
| "checksum": { |
| "kind": "sha256", |
| "value": "853f7ed317d4f2f80de646842e5bbc32f9d39c601562cbe466cda42e4bafb1b1" |
| }, |
| "purpose": "Paddle static graph learned parameters" |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 883, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "28dccd0baa8a939287aaa410cbacdbc94b2e0925" |
| }, |
| "purpose": "Paddle inference preprocessing/postprocessing config" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-onnx", |
| "repo": "PaddlePaddle/PP-OCRv6_tiny_det_onnx", |
| "revision": "2ba1506c0380b8f0b03dd142459aac66d4421f6c", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "inference.json", |
| "sizeBytes": 229184, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "56ddf4c08ad72e179d03a0caffdf794df5ffa47c" |
| } |
| }, |
| { |
| "path": "inference.onnx", |
| "sizeBytes": 1780590, |
| "checksum": { |
| "kind": "sha256", |
| "value": "193bab7a04fca699a6c82e6abb5b81bdb28177f0abd4062552b04908dafb19f8" |
| } |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 883, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "28dccd0baa8a939287aaa410cbacdbc94b2e0925" |
| } |
| } |
| ], |
| "notes": "Official ONNX sibling for regression checks; conversion source remains the Paddle static repo." |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "ppocr_v6_tiny/PP-OCRv6_tiny_det_onnx_infer.tar", |
| "sourceUrl": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-OCRv6_tiny_det_onnx_infer.tar", |
| "sha256": "ff6ab415b0a6e0c488550f2fb5d5046f1719848df220b2dc21b56402a65bc05d", |
| "sizeBytes": 1792000, |
| "format": "tar", |
| "notes": "Current PaddleOCR.js archive used by the project runtime." |
| } |
| ], |
| "conversionNotes": [ |
| "Use the Paddle static source files as the jax-js conversion source of truth.", |
| "Use official ONNX sibling and current PaddleOCR.js tar only as regression oracles." |
| ] |
| }, |
| { |
| "modelId": "ppocr-v6-medium-det", |
| "role": "ocr.det", |
| "family": "PP-OCRv6_medium_det", |
| "upstreamModelFamily": "PP-OCRv6_medium_det", |
| "source": { |
| "kind": "huggingface-paddle-static", |
| "repo": "PaddlePaddle/PP-OCRv6_medium_det", |
| "revision": "8e0f56fb2ef86b461d99cfc7ac5c137738985f61", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "inference.json", |
| "sizeBytes": 312150, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "548f60c00e9303a0323feceeae460256576671a6" |
| }, |
| "purpose": "Paddle static graph program description" |
| }, |
| { |
| "path": "inference.pdiparams", |
| "sizeBytes": 61960476, |
| "checksum": { |
| "kind": "sha256", |
| "value": "85218d2e3d98f5a21c58b4220627be923a97aee5db3cc71f39536ab31ac53960" |
| }, |
| "purpose": "Paddle static graph learned parameters" |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 886, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "1c5c05809877e4c7385f899019fff0ac9017ca80" |
| }, |
| "purpose": "Paddle inference preprocessing/postprocessing config" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-onnx", |
| "repo": "PaddlePaddle/PP-OCRv6_medium_det_onnx", |
| "revision": "61323801669c338b7891481ec7bac61ce31b576a", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "inference.json", |
| "sizeBytes": 312150, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "548f60c00e9303a0323feceeae460256576671a6" |
| } |
| }, |
| { |
| "path": "inference.onnx", |
| "sizeBytes": 62032837, |
| "checksum": { |
| "kind": "sha256", |
| "value": "eb13b44b25bb36f89528b68720af8a61d9cf381176107f465db1757b65d086e1" |
| } |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 886, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "1c5c05809877e4c7385f899019fff0ac9017ca80" |
| } |
| } |
| ], |
| "notes": "Official ONNX sibling for regression checks; conversion source remains the Paddle static repo." |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "ppocr_v6_medium/inference.onnx", |
| "sourceUrl": "https://huggingface.co/PaddlePaddle/PP-OCRv6_medium_det_onnx/resolve/61323801669c338b7891481ec7bac61ce31b576a/inference.onnx", |
| "sha256": "eb13b44b25bb36f89528b68720af8a61d9cf381176107f465db1757b65d086e1", |
| "sizeBytes": 62032837, |
| "format": "onnx", |
| "sidecars": [ |
| { |
| "role": "config", |
| "localPath": "ppocr_v6_medium/inference.yml", |
| "sourceUrl": "https://huggingface.co/PaddlePaddle/PP-OCRv6_medium_det_onnx/resolve/61323801669c338b7891481ec7bac61ce31b576a/inference.yml", |
| "sha256": "7298d5ead546584af2504d03355f881ac7a7bc0eb1e282d3e159277c1d0af871", |
| "format": "yml" |
| } |
| ], |
| "notes": "Optional official ONNX artifact retained as the medium-det conversion oracle." |
| } |
| ], |
| "conversionNotes": [ |
| "Use the Paddle static source files as the jax-js conversion source of truth.", |
| "Use the official ONNX sibling only as a regression oracle." |
| ] |
| }, |
| { |
| "modelId": "ppocr-v6-medium-rec", |
| "role": "ocr.rec", |
| "family": "PP-OCRv6_medium_rec", |
| "upstreamModelFamily": "PP-OCRv6_medium_rec", |
| "source": { |
| "kind": "huggingface-paddle-static", |
| "repo": "PaddlePaddle/PP-OCRv6_medium_rec", |
| "revision": "e5a92bcbc5cc1b494628e458d267778f0704fd7c", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "inference.json", |
| "sizeBytes": 221814, |
| "checksum": { "kind": "gitBlobSha1", "value": "9e0192344de3b69dacdb19dd735efbd06bd852f9" }, |
| "purpose": "Paddle static graph program description" |
| }, |
| { |
| "path": "inference.pdiparams", |
| "sizeBytes": 76465087, |
| "checksum": { "kind": "sha256", "value": "1b01c79a914587933f615569e75de54f2e638ebb5d3f3b3c1b38c24ede8c7319" }, |
| "purpose": "Paddle static graph learned parameters" |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 150580, |
| "checksum": { "kind": "gitBlobSha1", "value": "c53a96fcd315a86cb4748d2746f3d90941e1c6d8" }, |
| "purpose": "Paddle inference preprocessing, label dictionary, and postprocessing config" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-onnx", |
| "repo": "PaddlePaddle/PP-OCRv6_medium_rec_onnx", |
| "revision": "50c7eacafc52fa7bcf4194e8cd08e46f8558504b", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "inference.json", |
| "sizeBytes": 221814, |
| "checksum": { "kind": "gitBlobSha1", "value": "9e0192344de3b69dacdb19dd735efbd06bd852f9" } |
| }, |
| { |
| "path": "inference.onnx", |
| "sizeBytes": 76554979, |
| "checksum": { "kind": "sha256", "value": "9c09abf0957f7968c7586464b7397b84ad2387a0497a351af40e9acc71b673ba" } |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 150580, |
| "checksum": { "kind": "gitBlobSha1", "value": "c53a96fcd315a86cb4748d2746f3d90941e1c6d8" } |
| } |
| ], |
| "notes": "Official ONNX sibling for regression checks; conversion source remains the Paddle static repo." |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "ppocr_v6_medium/inference-rec.onnx", |
| "sourceUrl": "https://huggingface.co/PaddlePaddle/PP-OCRv6_medium_rec_onnx/resolve/50c7eacafc52fa7bcf4194e8cd08e46f8558504b/inference.onnx", |
| "sha256": "9c09abf0957f7968c7586464b7397b84ad2387a0497a351af40e9acc71b673ba", |
| "sizeBytes": 76554979, |
| "format": "onnx", |
| "sidecars": [ |
| { |
| "role": "config", |
| "localPath": "ppocr_v6_medium/inference-rec.yml", |
| "sourceUrl": "https://huggingface.co/PaddlePaddle/PP-OCRv6_medium_rec_onnx/resolve/50c7eacafc52fa7bcf4194e8cd08e46f8558504b/inference.yml", |
| "sha256": "991b700facf5b50a7de193468207d5f4255b538dde0d312ae3b7c7a9b6873129", |
| "format": "yml" |
| } |
| ], |
| "notes": "Optional official ONNX artifact retained as the medium-rec conversion oracle." |
| } |
| ], |
| "conversionNotes": [ |
| "Use the Paddle static source files as the jax-js conversion source of truth.", |
| "Use the official ONNX sibling only as a regression oracle." |
| ] |
| }, |
| { |
| "modelId": "ppocr-v6-tiny-rec", |
| "role": "ocr.rec", |
| "family": "PP-OCRv6_tiny_rec", |
| "upstreamModelFamily": "PP-OCRv6_tiny_rec", |
| "source": { |
| "kind": "huggingface-paddle-static", |
| "repo": "PaddlePaddle/PP-OCRv6_tiny_rec", |
| "revision": "0736086f72f666350ebcdc0c3a504eeac89cdfad", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "inference.json", |
| "sizeBytes": 108959, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "a99e3694b20c1c6d584cf33850f69416fe83cce4" |
| }, |
| "purpose": "Paddle static graph program description" |
| }, |
| { |
| "path": "inference.pdiparams", |
| "sizeBytes": 4420410, |
| "checksum": { |
| "kind": "sha256", |
| "value": "bb2f8f54d1e25f28c71b6fa4fe23f5940e159cae27fbee96155c99f822156e57" |
| }, |
| "purpose": "Paddle static graph learned parameters" |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 55571, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "85905320b166ff9fd95efbb69d1b0107f8ea3150" |
| }, |
| "purpose": "Paddle inference preprocessing, dictionary, and decoder config" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-onnx", |
| "repo": "PaddlePaddle/PP-OCRv6_tiny_rec_onnx", |
| "revision": "2612ab37152ae0a677521bae4e1e3d4fb4cf7c30", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "inference.json", |
| "sizeBytes": 108959, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "a99e3694b20c1c6d584cf33850f69416fe83cce4" |
| } |
| }, |
| { |
| "path": "inference.onnx", |
| "sizeBytes": 4462639, |
| "checksum": { |
| "kind": "sha256", |
| "value": "9ef676d6ed3c88256a2d92c640c44f25b0c40947e111b14b8be8f594091563e6" |
| } |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 55571, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "85905320b166ff9fd95efbb69d1b0107f8ea3150" |
| } |
| } |
| ], |
| "notes": "Official ONNX sibling for regression checks; conversion source remains the Paddle static repo." |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "ppocr_v6_tiny/PP-OCRv6_tiny_rec_onnx_infer.tar", |
| "sourceUrl": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-OCRv6_tiny_rec_onnx_infer.tar", |
| "sha256": "1e13b22717b1edd89d4cde4fda272b6c17d5b505c97c2baea99da1a3a2d54b29", |
| "sizeBytes": 4526080, |
| "format": "tar", |
| "notes": "Current PaddleOCR.js archive used by the project runtime." |
| } |
| ], |
| "conversionNotes": [ |
| "Preserve recognition dictionary and decoder semantics from inference.yml.", |
| "Use official ONNX sibling and current PaddleOCR.js tar only as regression oracles." |
| ] |
| }, |
| { |
| "modelId": "pp-doclayout-l", |
| "role": "layout.det", |
| "family": "PP-DocLayout-L", |
| "upstreamModelFamily": "PP-DocLayout-L", |
| "source": { |
| "kind": "huggingface-paddle-static", |
| "repo": "PaddlePaddle/PP-DocLayout-L", |
| "revision": "ca760c3336d922ee63cdbff85003181222e3911c", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "config.json", |
| "sizeBytes": 6134, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "0a3507f7365b0554e4e86d436421c53c3a738fb1" |
| }, |
| "purpose": "Model family metadata and label configuration" |
| }, |
| { |
| "path": "inference.json", |
| "sizeBytes": 1081389, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "540c4f71da8b65c044f14a242314cc73de10d3d3" |
| }, |
| "purpose": "Paddle static graph program description" |
| }, |
| { |
| "path": "inference.pdiparams", |
| "sizeBytes": 129021913, |
| "checksum": { |
| "kind": "sha256", |
| "value": "4df69115349e1e6215629d058e3bc7f087cad7a0d2776ef40c463daf73c38982" |
| }, |
| "purpose": "Paddle static graph learned parameters" |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 1871, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "632a2724ad049bbb1ea75777597ad7f12d106692" |
| }, |
| "purpose": "Paddle inference preprocessing/postprocessing config" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-community-onnx", |
| "repo": "x3zvawq/paddleocr-js-onnx", |
| "revision": "51c2133b5a7ea27b795fa8c400fdbfbd5337dd6a", |
| "license": "unknown", |
| "files": [ |
| { |
| "path": "pp_doclayout_l/PP-DocLayout-L_infer.onnx", |
| "sizeBytes": 129398697, |
| "checksum": { |
| "kind": "sha256", |
| "value": "01fd1a44fbea5b0a76302de356c1518250cbd34ee82833ac04d907034c1376e1" |
| } |
| }, |
| { |
| "path": "pp_doclayout_l/PP-DocLayout-L_inference.yml", |
| "sizeBytes": 1871, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "632a2724ad049bbb1ea75777597ad7f12d106692" |
| } |
| } |
| ], |
| "notes": "Community ONNX artifact used by the current project as a layout regression oracle." |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "pp_doclayout_l/PP-DocLayout-L_infer.onnx", |
| "sourceUrl": "https://huggingface.co/x3zvawq/paddleocr-js-onnx/resolve/main/pp_doclayout_l/PP-DocLayout-L_infer.onnx", |
| "sha256": "01fd1a44fbea5b0a76302de356c1518250cbd34ee82833ac04d907034c1376e1", |
| "sizeBytes": 129398697, |
| "format": "onnx", |
| "sidecars": [ |
| { |
| "role": "config", |
| "localPath": "pp_doclayout_l/PP-DocLayout-L_inference.yml", |
| "sourceUrl": "https://huggingface.co/x3zvawq/paddleocr-js-onnx/resolve/main/pp_doclayout_l/PP-DocLayout-L_inference.yml", |
| "sha256": "fbdbb903efd3d82db5800f9ae3e2477d2d84525956ef410ffe8e64bbaad02fa5", |
| "format": "yml" |
| } |
| ], |
| "notes": "Current ONNX Runtime Web artifact used by the project runtime." |
| } |
| ], |
| "conversionNotes": [ |
| "Use static Paddle source as the conversion source of truth.", |
| "Treat the x3zvawq ONNX files as regression oracles, not as jax-js conversion input." |
| ] |
| }, |
| { |
| "modelId": "pp-doclayout-v3", |
| "role": "layout.det", |
| "family": "PP-DocLayoutV3", |
| "upstreamModelFamily": "PP-DocLayoutV3", |
| "source": { |
| "kind": "huggingface-paddle-static", |
| "repo": "PaddlePaddle/PP-DocLayoutV3", |
| "revision": "7b48a7566925fa464281f930c58eee04fe2c862a", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "README.md", |
| "sizeBytes": 12077, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "0e67905a060d65fffd43d158c078e320b3c5d784" |
| }, |
| "purpose": "Pinned official model card, Apache-2.0 license metadata, and usage contract" |
| }, |
| { |
| "path": "inference.json", |
| "sizeBytes": 1196890, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "5fd3ae9d9615dd36cd2134551a3bcc60d222ed21" |
| }, |
| "purpose": "Paddle static graph program description" |
| }, |
| { |
| "path": "inference.pdiparams", |
| "sizeBytes": 130806572, |
| "checksum": { |
| "kind": "sha256", |
| "value": "70bd316b0582769ec968829fd1feb1a6a58b7c941b938327e551b6b12b45c137" |
| }, |
| "purpose": "Paddle static graph learned parameters" |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 1482, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "ed7472400b398e0e0e032893f7986b32692980e7" |
| }, |
| "purpose": "Paddle preprocessing, 25-class label mapping, threshold, and layout postprocessing config" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-safetensors", |
| "repo": "PaddlePaddle/PP-DocLayoutV3_safetensors", |
| "revision": "97d101e6db2642e162a1d05392d1b0231c91033e", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "README.md", |
| "sizeBytes": 13147, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "48fc6ecdb0cec7cc38d759cdea3caaccba39ae4a" |
| }, |
| "purpose": "Official Transformers usage and decoded polygon/order output contract" |
| }, |
| { |
| "path": "config.json", |
| "sizeBytes": 2460, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "5a22928c191950850cbc0e56e43f722073e7c8da" |
| }, |
| "purpose": "Transformers architecture and numeric class configuration" |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 1482, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "ed7472400b398e0e0e032893f7986b32692980e7" |
| }, |
| "purpose": "Canonical static 25-class label mapping retained beside converted weights" |
| }, |
| { |
| "path": "model.safetensors", |
| "sizeBytes": 133270468, |
| "checksum": { |
| "kind": "sha256", |
| "value": "5ea422c6cc5fe759a47e1357c35639b58173508e025a3131cbe4b6ac59e2b85e" |
| }, |
| "purpose": "Official 33.3M-parameter Transformers conversion used as an independent oracle" |
| }, |
| { |
| "path": "preprocessor_config.json", |
| "sizeBytes": 575, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "ab66797648e5a3247eca2988e9fcd8af07a6a038" |
| }, |
| "purpose": "Official image preprocessing and postprocessing configuration" |
| } |
| ], |
| "notes": "Official independent weight/config oracle. Preserve numeric IDs and use inference.yml labels because config.json collapses several display names." |
| } |
| ], |
| "conversionNotes": [ |
| "Pin both official repositories; never convert mutable main.", |
| "Use Paddle static inference.json and inference.pdiparams as the primary JAXJS conversion source.", |
| "Use the official safetensors implementation as an independent decoded-output and architecture oracle.", |
| "Preserve numeric class IDs and inference.yml labels, including distinct display_formula and inline_formula.", |
| "Treat polygon masks and learned logical order as required outputs, not optional metadata." |
| ] |
| }, |
| { |
| "modelId": "pp-formulanet-plus-s", |
| "role": "formula.rec", |
| "family": "PP-FormulaNet_plus-S", |
| "upstreamModelFamily": "PP-FormulaNet_plus-S", |
| "source": { |
| "kind": "huggingface-paddle-static", |
| "repo": "PaddlePaddle/PP-FormulaNet_plus-S", |
| "revision": "3d46f557e3a1752f4bf81202395af3b5ecfadfd2", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "config.json", |
| "sizeBytes": 3951676, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "7984e53d42a00727656d7ba8e5292d5a1af59a05" |
| }, |
| "purpose": "Tokenizer and decoder configuration" |
| }, |
| { |
| "path": "inference.json", |
| "sizeBytes": 506956, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "d7709046268d4cff05deed68af33f977e9b626e2" |
| }, |
| "purpose": "Paddle static graph program description" |
| }, |
| { |
| "path": "inference.pdiparams", |
| "sizeBytes": 256845006, |
| "checksum": { |
| "kind": "sha256", |
| "value": "e464f94412feaa98f8791eacc84684f887b3569e30e80c52b8112e9cf7d4069b" |
| }, |
| "purpose": "Paddle static graph learned parameters" |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 2244564, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "8b422c78832435a715f50f106130ee97ca68b8e7" |
| }, |
| "purpose": "Paddle inference preprocessing/postprocessing config" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-community-onnx", |
| "repo": "x3zvawq/paddleocr-js-onnx", |
| "revision": "51c2133b5a7ea27b795fa8c400fdbfbd5337dd6a", |
| "license": "unknown", |
| "files": [ |
| { |
| "path": "PP_FormulaNet_plus_S/PP-FormulaNet_plus-S_infer.onnx", |
| "sizeBytes": 333913284, |
| "checksum": { |
| "kind": "sha256", |
| "value": "713d472d8299abbf936e0c7ac74de9c63cc62f9bcf2a01be5fa376bc1d30fb76" |
| } |
| }, |
| { |
| "path": "PP_FormulaNet_plus_S/config.json", |
| "sizeBytes": 3951676, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "7984e53d42a00727656d7ba8e5292d5a1af59a05" |
| } |
| }, |
| { |
| "path": "PP_FormulaNet_plus_S/inference.yml", |
| "sizeBytes": 2244564, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "8b422c78832435a715f50f106130ee97ca68b8e7" |
| } |
| } |
| ], |
| "notes": "Community ONNX artifact used by the current project as a formula regression oracle." |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "PP_FormulaNet_plus_S/PP-FormulaNet_plus-S_infer.onnx", |
| "sourceUrl": "https://huggingface.co/x3zvawq/paddleocr-js-onnx/resolve/main/PP_FormulaNet_plus_S/PP-FormulaNet_plus-S_infer.onnx", |
| "sha256": "713d472d8299abbf936e0c7ac74de9c63cc62f9bcf2a01be5fa376bc1d30fb76", |
| "sizeBytes": 333913284, |
| "format": "onnx", |
| "sidecars": [ |
| { |
| "role": "config", |
| "localPath": "PP_FormulaNet_plus_S/config.json", |
| "sourceUrl": "https://huggingface.co/x3zvawq/paddleocr-js-onnx/resolve/main/PP_FormulaNet_plus_S/config.json", |
| "sha256": "ddf1f951ceeb10c2b9de3ae255b62de59299b1028f87878177e86c9604fe3610", |
| "format": "json" |
| } |
| ], |
| "notes": "Current ONNX Runtime Web artifact used by the project runtime." |
| } |
| ], |
| "conversionNotes": [ |
| "Treat config.json as conversion-critical tokenizer/decoder metadata.", |
| "Use static Paddle source as conversion input and x3zvawq ONNX files as regression oracles." |
| ] |
| }, |
| { |
| "modelId": "slanet-plus-table", |
| "role": "table.rec", |
| "family": "SLANet-plus", |
| "upstreamModelFamily": "SLANet_plus", |
| "source": { |
| "kind": "huggingface-paddle-static", |
| "repo": "PaddlePaddle/SLANet_plus", |
| "revision": "bae6e5f8c3c4e7da0c0b7639fdf3228fe76184e2", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "config.json", |
| "sizeBytes": 4897, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "8933a73b813b20f2eb9d425da0c50142d0553c9c" |
| }, |
| "purpose": "Model family metadata" |
| }, |
| { |
| "path": "inference.json", |
| "sizeBytes": 315660, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "3cac190b2282531db445059f869054f37784ff90" |
| }, |
| "purpose": "Paddle static graph program description" |
| }, |
| { |
| "path": "inference.pdiparams", |
| "sizeBytes": 7666515, |
| "checksum": { |
| "kind": "sha256", |
| "value": "012986b0e2bfe90410618bfb4b7cf4fcb6c978caefd84c63c2934f8387b09ab8" |
| }, |
| "purpose": "Paddle static graph learned parameters" |
| }, |
| { |
| "path": "inference.yml", |
| "sizeBytes": 2035, |
| "checksum": { |
| "kind": "gitBlobSha1", |
| "value": "9d57e936f02d62335817f936d6e779e811f56d2a" |
| }, |
| "purpose": "Paddle inference preprocessing/postprocessing config" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-community-onnx", |
| "repo": "xberg-io/paddleocr-onnx-models", |
| "revision": "152b6c4ac9294aade646439baa38f8462f17012a", |
| "license": "apache-2.0", |
| "files": [ |
| { |
| "path": "v2/table/SLANet_plus.onnx", |
| "sizeBytes": 7781309, |
| "checksum": { |
| "kind": "sha256", |
| "value": "e48a401a4ebcddd47fe3822427db24d867a557324f58e438692f588bbe9231de" |
| } |
| } |
| ], |
| "notes": "Community ONNX artifact used by the current project as a table regression oracle." |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "v2/table/SLANet_plus.onnx", |
| "sourceUrl": "https://huggingface.co/xberg-io/paddleocr-onnx-models/resolve/main/v2/table/SLANet_plus.onnx", |
| "sha256": "e48a401a4ebcddd47fe3822427db24d867a557324f58e438692f588bbe9231de", |
| "sizeBytes": 7781309, |
| "format": "onnx", |
| "sidecars": [ |
| { |
| "role": "dict", |
| "localPath": "v2/table/table_structure_dict_ch.txt", |
| "sourceUrl": "https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/main/ppocr/utils/dict/table_structure_dict_ch.txt", |
| "sha256": "68d344a84b726e043f390122240ff2b2ced2949b2a80ce9b61ae955054d190ef", |
| "format": "txt" |
| } |
| ], |
| "notes": "Current ONNX Runtime Web artifact and table dictionary used by the project runtime." |
| } |
| ], |
| "conversionNotes": [ |
| "Preserve table structure dictionary semantics during jax-js decoding.", |
| "Use Paddle static source as conversion input and xberg ONNX file as a regression oracle." |
| ] |
| }, |
| { |
| "modelId": "mineru-table-classifier", |
| "role": "table.cls", |
| "family": "PP-LCNet_x1_0_table_cls", |
| "upstreamModelFamily": "PP-LCNet_x1_0_table_cls", |
| "source": { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabCls/paddle_table_cls/PP-LCNet_x1_0_table_cls.onnx", |
| "sizeBytes": 6776877, |
| "checksum": { "kind": "sha256", "value": "c84bf1d79c1c74d534b5b12adb14dd12151c42f7ae3e4be4f1042b830f80b949" }, |
| "purpose": "MinerU wired/wireless table classifier" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabCls/paddle_table_cls/PP-LCNet_x1_0_table_cls.onnx", |
| "sizeBytes": 6776877, |
| "checksum": { "kind": "sha256", "value": "c84bf1d79c1c74d534b5b12adb14dd12151c42f7ae3e4be4f1042b830f80b949" } |
| } |
| ], |
| "notes": "Exact CPU oracle checkpoint used by MinerU 3.4.2." |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "table/paddle_table_cls/PP-LCNet_x1_0_table_cls.onnx", |
| "sourceUrl": "https://huggingface.co/opendatalab/PDF-Extract-Kit-1.0/resolve/ed6b654c018d742e65a17671e379c5e6ecc87ec9/models/TabCls/paddle_table_cls/PP-LCNet_x1_0_table_cls.onnx", |
| "sha256": "c84bf1d79c1c74d534b5b12adb14dd12151c42f7ae3e4be4f1042b830f80b949", |
| "sizeBytes": 6776877, |
| "format": "onnx", |
| "notes": "Shared Node/browser ONNX Runtime artifact." |
| } |
| ], |
| "conversionNotes": [ |
| "Use a shared ONNX Runtime adapter because the source artifact is already a small pinned ONNX graph.", |
| "Keep preprocessing and label decoding in shared TypeScript and verify logits against the CPU oracle." |
| ] |
| }, |
| { |
| "modelId": "mineru-slanet-plus-table", |
| "role": "table.rec", |
| "family": "SLANet-plus-MinerU", |
| "upstreamModelFamily": "SLANetPlus", |
| "source": { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabRec/SlanetPlus/slanet-plus.onnx", |
| "sizeBytes": 7758305, |
| "checksum": { "kind": "sha256", "value": "d57a942af6a2f57d6a4a0372573c696a2379bf5857c45e2ac69993f3b334514b" }, |
| "purpose": "MinerU-selected wireless table recognizer" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabRec/SlanetPlus/slanet-plus.onnx", |
| "sizeBytes": 7758305, |
| "checksum": { "kind": "sha256", "value": "d57a942af6a2f57d6a4a0372573c696a2379bf5857c45e2ac69993f3b334514b" } |
| } |
| ] |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "table/slanet_plus/slanet-plus.onnx", |
| "sourceUrl": "https://huggingface.co/opendatalab/PDF-Extract-Kit-1.0/resolve/ed6b654c018d742e65a17671e379c5e6ecc87ec9/models/TabRec/SlanetPlus/slanet-plus.onnx", |
| "sha256": "d57a942af6a2f57d6a4a0372573c696a2379bf5857c45e2ac69993f3b334514b", |
| "sizeBytes": 7758305, |
| "format": "onnx", |
| "sidecars": [ |
| { |
| "role": "dict", |
| "localPath": "table/slanet_plus/table_structure_dict_ch.txt", |
| "sourceUrl": "https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/main/ppocr/utils/dict/table_structure_dict_ch.txt", |
| "sha256": "68d344a84b726e043f390122240ff2b2ced2949b2a80ce9b61ae955054d190ef", |
| "format": "txt" |
| } |
| ] |
| } |
| ], |
| "conversionNotes": [ |
| "The bytes differ from the existing xberg SLANet artifact, so expose an explicit MinerU candidate instead of aliasing it.", |
| "Reuse the shared table decoder only after tensor/semantic oracle parity is demonstrated." |
| ] |
| }, |
| { |
| "modelId": "mineru-slanet-plus-table-ort-web", |
| "role": "table.rec", |
| "family": "SLANet-plus-MinerU", |
| "upstreamModelFamily": "SLANetPlus", |
| "source": { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabRec/SlanetPlus/slanet-plus.onnx", |
| "sizeBytes": 7758305, |
| "checksum": { "kind": "sha256", "value": "d57a942af6a2f57d6a4a0372573c696a2379bf5857c45e2ac69993f3b334514b" }, |
| "purpose": "Parent weights for the metadata-only ONNX Runtime Web repair" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabRec/SlanetPlus/slanet-plus.onnx", |
| "sizeBytes": 7758305, |
| "checksum": { "kind": "sha256", "value": "d57a942af6a2f57d6a4a0372573c696a2379bf5857c45e2ac69993f3b334514b" } |
| } |
| ] |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "table/slanet_plus/slanet-plus-ort-web.onnx", |
| "sourceUrl": "https://huggingface.co/hf-dev-01/pj-ocr-jaxjs-medium/resolve/3f79d3558a5903916732bfe9d5166c217832f20d/models/table/slanet_plus/slanet-plus-ort-web.onnx", |
| "sha256": "11b311f67b2862e96476bd46318cd0ba4b75e4b629901ba94f020d74e4d8e532", |
| "sizeBytes": 7758309, |
| "format": "onnx", |
| "sidecars": [ |
| { |
| "role": "dict", |
| "localPath": "table/slanet_plus/table_structure_dict_ch.txt", |
| "sourceUrl": "https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/main/ppocr/utils/dict/table_structure_dict_ch.txt", |
| "sha256": "68d344a84b726e043f390122240ff2b2ced2949b2a80ce9b61ae955054d190ef", |
| "format": "txt" |
| } |
| ], |
| "notes": "Metadata-only scalar-to-[1] repair; frozen native ORT outputs remain byte-identical." |
| } |
| ], |
| "conversionNotes": [ |
| "Apply only the reproducible scalar-to-[1] rank metadata repair required by ONNX Runtime Web 1.27.", |
| "Require byte-identical native ORT outputs on both pinned paper-table crops." |
| ] |
| }, |
| { |
| "modelId": "mineru-unet-structure-table", |
| "role": "table.wired", |
| "family": "UnetStructure", |
| "upstreamModelFamily": "UnetStructure", |
| "source": { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabRec/UnetStructure/unet.onnx", |
| "sizeBytes": 8335007, |
| "checksum": { "kind": "sha256", "value": "0ea48d3a17e35ef5c2e498a5e799566073234d39b1079ca21d9f4fafe73c6d20" }, |
| "purpose": "MinerU wired table line-segmentation model" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabRec/UnetStructure/unet.onnx", |
| "sizeBytes": 8335007, |
| "checksum": { "kind": "sha256", "value": "0ea48d3a17e35ef5c2e498a5e799566073234d39b1079ca21d9f4fafe73c6d20" } |
| } |
| ] |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "table/unet_structure/unet.onnx", |
| "sourceUrl": "https://huggingface.co/opendatalab/PDF-Extract-Kit-1.0/resolve/ed6b654c018d742e65a17671e379c5e6ecc87ec9/models/TabRec/UnetStructure/unet.onnx", |
| "sha256": "0ea48d3a17e35ef5c2e498a5e799566073234d39b1079ca21d9f4fafe73c6d20", |
| "sizeBytes": 8335007, |
| "format": "onnx" |
| } |
| ], |
| "conversionNotes": [ |
| "Use the small pinned ONNX graph through the shared runtime boundary; keep OpenCV-style line/cell recovery as explicit deterministic postprocessing.", |
| "Verify segmentation tensors and decoded cell polygons independently before pipeline promotion." |
| ] |
| }, |
| { |
| "modelId": "mineru-unet-structure-table-fp16", |
| "role": "table.wired", |
| "family": "UnetStructure", |
| "upstreamModelFamily": "UnetStructure", |
| "source": { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabRec/UnetStructure/unet.onnx", |
| "sizeBytes": 8335007, |
| "checksum": { "kind": "sha256", "value": "0ea48d3a17e35ef5c2e498a5e799566073234d39b1079ca21d9f4fafe73c6d20" }, |
| "purpose": "FP32 oracle parent for the approved FP16-storage artifact" |
| } |
| ] |
| }, |
| "oracles": [ |
| { |
| "kind": "huggingface-onnx", |
| "repo": "opendatalab/PDF-Extract-Kit-1.0", |
| "revision": "ed6b654c018d742e65a17671e379c5e6ecc87ec9", |
| "license": "AGPL-3.0", |
| "files": [ |
| { |
| "path": "models/TabRec/UnetStructure/unet.onnx", |
| "sizeBytes": 8335007, |
| "checksum": { "kind": "sha256", "value": "0ea48d3a17e35ef5c2e498a5e799566073234d39b1079ca21d9f4fafe73c6d20" } |
| } |
| ] |
| }, |
| { |
| "kind": "manifest-artifact", |
| "localPath": "table/unet_structure/unet-fp16.onnx", |
| "sourceUrl": "https://huggingface.co/hf-dev-01/pj-ocr-jaxjs-medium/resolve/3f79d3558a5903916732bfe9d5166c217832f20d/models/table/unet_structure/unet-fp16.onnx", |
| "sha256": "2ad99d71a6873284ebd414d1cbdf5c4ebead299ccc484087988d5832c30803f8", |
| "sizeBytes": 4197486, |
| "format": "onnx", |
| "notes": "R90AA10-023-selected FP16-storage derivative with non-vacuous structure/HTML parity." |
| } |
| ], |
| "conversionNotes": [ |
| "Convert floating initializers reproducibly while preserving operator-required dtypes.", |
| "Approve only after segmentation, recovered grid/cells, HTML, arbitration, repetition, latency, and RSS gates pass." |
| ] |
| } |
| ] |
| } |
|
|
