OCRia-1

Headlines · 2

1. Nov 2
   LlamaIndex reports nearly 10x faster token generation for OCRia-1 on one B200

   LlamaIndex reports that its optimizations raised OCRia-1 token generation from about 285 to 2,800 tokens per second on a single NVIDIA B200 GPU, a vendor-reported result not independently verified here. The stack combines workload-aware mixed-precision quantization, DFlash2 speculative decoding, ReplaySSM, custom kernels and 17 additional optimizations. LlamaIndex says OCRia-1 is available through LlamaParse across all tiers and plans to contribute generally applicable improvements to vLLM.

   LlamaIndex Blog · 🔥 0
2. Oct 22
   LlamaIndex launches 4B document parsing model OCRia-1

   LlamaIndex introduced OCRia-1, a 4-billion-parameter vision-language model designed to reconstruct complex document tables, charts, formatting and mathematical notation. The developer reports that it scored 87.4 on ParseBench and led table evaluations on olmOCR-bench, PulseBench-Tab and fr-bench-pdf2md, and the model is now integrated into LlamaParse without a pricing change.

   LlamaIndex Blog · 🔥 0

Experience and discussion from the community

Share my OCRia-1 experienceAsk about OCRia-1