Datasetpublic

allenai/olmOCR-mix-0225-benchmarkset

This is just 10,000 PDFs randomly sampled from https://huggingface.co/datasets/allenai/olmOCR-mix-0225 that we use internally at AI2 for speed benchmarking and also quantization calibration.

sourceHugging Faceodc-byupdated 1y agoView on Hugging Face

2likes447downloads

Dataset Card

This is just 10,000 PDFs randomly sampled from https://huggingface.co/datasets/allenai/olmOCR-mix-0225 that we use internally at AI2 for speed benchmarking and also quantization calibration.