Datalab ships Indra OCR 1.1: 1M-param model does 44 pages/s on CPU, near DeepSeek OCR 2 accuracy
VikParuchuri · x · 2026-10-20
Datalab (VikParuchuri) released Indra OCR 1.1, a tiny-but-fast document OCR model: a 1M-parameter main model running 44 pages/s on 32 CPU cores (15x faster than Tesseract and more accurate, 10x PaddleOCR's PP-OCR text models, 210ms/page on 4 cores). It scores 66.7 on OmniParseBench (71.4 with --math), ahead of PaddleOCR's full PP-StructureV3 pipeline and within 4 points of the 3.4B-param DeepSeek OCR 2. Supports 80+ languages plus handwriting, handles scans tilted up to 45°, outputs per-line boxes and reading order, and runs on ONNX Runtime with no PyTorch or GPU. Weights are open.
Related event: Datalab Releases Indra OCR 1.1: 1M-Parameter Model Runs Fast on CPU(3 posts)→
More from Infra
- Report: Liang Wenfeng planned independent DeepSeek compute chips before Huawei persuaded him otherwise — teortaxesTex · 2026-10-20
- TwelveLabs Pegasus 1.5 lands on Amazon Bedrock: turn 2-hour videos into timestamped JSON — AWS ML Blog · 2026-10-20
- NVIDIA's Ian Buck explains how AI factories can flex power use to relieve the grid — nvidia · 2026-10-20
- p99 container startup latency cut 63%, saving ~19 hours of compute daily on Cloud Run — DanielLockyer · 2026-10-20
- Late Interaction in 2026: Experiment Shows Single-Vector Search Ranks Wrong Answers Higher — IgorCarron · 2026-10-20
- AMD, NVIDIA, and Google All Ship Major vLLM Inference Optimizations in Three Days — PyTorch · 2026-10-20