Skip to content

Added information and error message for Python 3.14 users facing mark…#2

Open
havelentzas wants to merge 1 commit into
scdenney:mainfrom
havelentzas:main
Open

Added information and error message for Python 3.14 users facing mark…#2
havelentzas wants to merge 1 commit into
scdenney:mainfrom
havelentzas:main

Conversation

@havelentzas

Copy link
Copy Markdown

Python 3.14 compatibility warning for PDF input

Summary

  • Updated README to document that PDF input does not work on Python 3.14; .md, .markdown, .txt, and .tex input works fine
  • Updated pyproject.toml classifiers to include Python 3.13
  • Added a clear error message in pipeline.py that surfaces when marker-pdf fails on Python 3.14, directing users to the README and suggesting they pass a .md file or downgrade to Python 3.12/3.13

Background

Also see #1

marker-pdf depends on surya-ocr 0.12.x, which was built against transformers 4.43.x. On Python 3.14, tokenizers 0.20.x (required to downgrade transformers) cannot be compiled from source because pyo3 0.22.5 only supports up to Python 3.13. The transformers version that installs in a Python 3.14 environment (4.57+) has breaking changes for surya's composition config classes that cause PDF conversion to fail. Passing a pre-converted .md file bypasses marker entirely and works without issue on Python 3.14.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant