nielsr/arxiv-chandra-ocr-smoke-20260328-tokenfix

arXiv OCR with Chandra OCR 2

This dataset stores OCR results for arXiv PDFs using datalab-to/chandra-ocr-2.

Summary

• —Output dataset: nielsr/arxiv-chandra-ocr-smoke-20260328-tokenfix
• —Source paper IDs in input list: 2
• —Processed IDs recorded in state/processed_ids.txt: 2
• —Successes: 2
• —Partial successes: 0
• —Errors: 0
• —Next shard index: 2
• —Updated at: 2026-03-28T15:22:09.273986+00:00

Files

• —data/part-*.jsonl.gz: OCR result shards, one JSON object per paper
• —state/processed_ids.txt: completed paper IDs used for resume
• —state/summary.json: aggregate counters and bookkeeping

Load the results

from datasets import load_dataset

dataset = load_dataset("nielsr/arxiv-chandra-ocr-smoke-20260328-tokenfix", data_files="data/*.jsonl.gz", split="train")
print(dataset[0]["paper_id"])
print(dataset[0]["markdown"][:1000])

Job config

• —Prompt type: ocr_layout
• —Page batch size: 16
• —Max output tokens: 12384
• —Max model length: 18000
• —GPU memory utilization: 0.85
• —Minimum arXiv request interval: 3.1 seconds
• —Paginate output: False
• —Include headers/footers: False

Reproduction

hf jobs uv run --flavor l4x1 --image vllm/vllm-openai:v0.17.0 \
  -s HF_TOKEN --timeout 2d \
  ./chandra2-arxiv-ocr.py nielsr/arxiv-chandra-ocr-smoke-20260328-tokenfix \
  --paper-ids-url https://.../hf_missing_paper_ids.txt