DeepSeek-OCR-2 on every GPU you own — vLLM or transformers on CUDA, native Apple Silicon, and CPU, with per-page quality checks you can trust. Added - New backend package s

- Chuyển sang Model Deepseek-OCR-2 + transformers - Thêm hỗ trợ cho Linux

…0.4.23 . After upgrading to v0.4.23 , fine-tuning DeepSeek-OCR (or any VLM with a custom processor that carries image data under a non-standard key) would crash at inferenc

…-to-video.ipynb) - [Post-Training Optimization of DeepSeek-OCR Model](https://github.com/openvinotoolkit/openvino notebooks/blob/latest/notebooks/deepseek-ocr/deepseek-ocr.

…ode, we’re ready to introduce DSQ quantization to deepseek-ocr.rs . Unlike formats like GGUF or plain safetensors that usually ship full float checkpoints, DSQ behaves as a

…ons on Blackwell GPUs https://github.com/TimmyOVO/deepseek-ocr.rs/issues/24 Blackwell builds were blowing up memory because the RoPE cache tried to grow to ~4e18 positions.