重點導讀
Unlimited OCR 2 已推出,並已開源。
- - SR‑SWA 機制:將文字‑視覺注意力互動壓縮為固定大小的工作參照,降低長檔案計算負擔。
- - 長檔案解析:支援任意長度檔案,且隨檔案增長保持近乎恆定的逐步解碼延遲與記憶體佔用。
- - 基準表現:在 MPDocBench 與 MarathonDocBench 上達到領先水平,涵蓋數十至數百頁檔案。
AI 伴隨問答
針對這篇報導有疑問?可直接點選快捷問題或向 AI 追問背景脈絡與技術細節:
Baidu Inc.@Baidu_Inc
·27 天前
Unlimited OCR 2 is here, and it's open source. 📖 ✨
Longer documents can mean a lot more attention computation. Our new SR-SWA mechanism keeps that in check by condensing text-visual attention interactions into a fixed-size set of working references.
Here's what it delivers:
- Document parsing at virtually any length, within a finite context window
- Nearly constant per-step decoding latency and memory footprint as documents grow longer
- State-of-the-art performance on MPDocBench and MarathonDocBench, covering documents from dozens to hundreds of pages
Check out the model here ↓
推文來源 · X (Twitter)在 X 參與討論
來源: Baidu Inc. (x.com)
在 X 查看推文