Skip to content

MiniCPM-OCR SDK v1.0.4

Latest
Latest

Choose a tag to compare

View all tags
@huangfw huangfw released this 03 Nov 02:50
· 10 commits to main since this release
sdk-v1.0.4
f7d893a
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Page output is aligned with the online document parsing API, and adaptive rotation is enabled by default.

  • Online isomorphic results: for each page, an additional pages_json/<image_id>.json is written, with a structure consistent with the single-page response of the online POST /v1/document-parse (metadata.page_index, choices[0].message.content, ocr.blocks, ocr.rotation, usage); minicpmvocr parse --json FILE and assemble_document_json() are added, producing the full result consistent with the online asynchronous task.

  • Unified block statuses: only ok, failed, skipped_no_vlm, and skipped_seal are used; recognition failures and missing crops are uniformly recorded as failed, with the specific reason retained in the run log; when seal recognition is disabled, seal blocks are recorded as skipped_seal.

  • Rotation logging: ocr.rotation is consistent with the online field, including skipping for small images (gated) and unavailable when paddleocr is missing.

  • Adaptive rotation is enabled by default: Quick Start automatically downloads the three Paddle orientation-correction models; when paddleocr is not installed, a warning is logged and orientation correction is skipped instead of raising an error. To disable it, use --set upright.enabled=false; disabling is recommended when reproducing OmniDocBench scores.

  • GIF input is supported.

Assets 2
Loading