GitHub - baidu/Unlimited-OCR: Unlimited OCR Works: Welcome the Era of One-shot Long-horizon Parsing.

kept by eddie jun 24

Baidu releases Unlimited-OCR, a VLM-based OCR that processes entire documents in a single pass using reference sliding window attention to avoid memory overflow.

Review of Deep Seek OCR
DeepSeek-OCR compresses image tokens using an encoder to reduce input size, enabling larger context windows and more efficient training, not traditional OCR.
also on DeepSeek-OCR
How China is getting everyone on OpenClaw, from gearheads to grandmas
China's tech giants are hosting public events to drive mass adoption of OpenClaw, the AI agent now used beyond the U.S. and supported by government subsidies.
also on Baidu, #open-source
Advancements In Self-Driving Cars - by Zvi Mowshowitz
Waymo expands to the entire SF Bay Area and plans 17 US cities by 2026, but faces regulatory barriers and speed-limit issues despite being vastly safer than human drivers.
also on Baidu
Why Open-Source Models Haven't Killed the Big Dogs
Open-source models remain costlier to serve reliably due to infrastructure, utilization, and operational overhead, making hosted APIs from OpenAI and Anthropic more economical despite free model weights.
also on SGLang, #generative-ai, #ai-integration
[2606.26907] Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation
The paper proposes Qwen-Image-Agent, a unified agentic framework that uses planning, reasoning, search, memory, and feedback to bridge the context gap in real-world text-to-image generation.
also on #computer-vision, #generative-ai