DOC//OCR
PP-DocLayout + PaddleOCR-VL - Powered By BWS
→ AUTH_KEY_REQUIRED:
[SELECT FILE]
// No file selected...
USE vLLM BACKEND (Docker/GPU server)
USE LLAMA.CPP BACKEND (Fastest via genai_server)
USE LOCAL PADDLE VLM (Pure doc_parser / No Server)
RUN → PARSE
╔ OUTPUT ╗
-
Pages
-
Words
[RENDERED]
[RAW.MD]
COPY