flash-attn @ https://github.com/Dao-AILab/flash-attention/releases/download/v2.7.3/flash_attn-2.7.3+cu12torch2.6cxx11abiFALSE-cp310-cp310-linux_x86_64.whl transformers-stream-generator huggingface_hub qwen-vl-utils sentencepiece opencv-python torch==2.6.0 docling-core transformers torchvision matplotlib accelerate requests hf_xet spaces pillow gradio einops peft fpdf timm av