百度开源了Unlimited-OCR，在vLLM上跑，能一次性解析整本书，内存不涨，比DeepSeek-OCR快35%，做文档OCR的好东西。
百度Unlimited-OCR现已集成到vLLM推理框架中，基于Reference Sliding Window Attention（R-SWA）机制实现恒定KV缓存，避免内存暴涨和速度下降。该模型能在32K上下文预算下一次性转录40+页文档，且编辑距离极低。在6K输出token场景下，推理速度比DeepSeek-OCR快35%，GPU内存和吞吐量保持恒定。
🎉 Unlimited-OCR from @Baidu_Inc now runs in vLLM. …
🎉 Unlimited-OCR from @Baidu_Inc now runs in vLLM. One-shot parsing of entire books with constant KV cache, powered by Reference Sliding Window Attention (R-SWA).
🧠 R-SWA keeps KV cache fixed throughout decoding — no memory blowup, no slowdown, no matter how long the output gets. 📄 Transcribe 40+ pages in a single forward pass under a 32K context budget, with remarkably low edit distance even at scale. 🪶 35% faster than DeepSeek-OCR at 6K output tokens, with fully constant TPS and GPU memory.
🔗 Recipe: https://t.co/gBAvAP7vyG 🤗 Weights: https://t.co/9hnfOt3Ab7
🙏 Thanks to the @BaiduAI_News team for the collaboration.
- AK06-28 18:37原文