DeepSeek-OCR-Colab / requirements.txt
ahczhg's picture
Upload requirements.txt with huggingface_hub
09393ef verified
# DeepSeek-OCR Colab Requirements
# For local setup only - Google Colab already includes most of these
# Core dependencies
torch>=2.0.0
torchvision>=0.15.0
torchaudio>=2.0.0
# Transformers and tokenizers
transformers==4.46.3
tokenizers==0.20.3
# PDF and image processing
PyMuPDF>=1.23.0
img2pdf>=0.5.0
Pillow>=10.0.0
# Utilities
einops>=0.7.0
easydict>=1.10
addict>=2.4.0
numpy>=1.24.0
# Flash Attention (optional, may require compilation)
# flash-attn==2.7.3
# Additional dependencies
safetensors>=0.4.1
huggingface-hub>=0.23.2
# For Jupyter notebooks (local)
jupyter>=1.0.0
ipywidgets>=8.0.0
# Note: Flash Attention installation can be challenging and may require:
# - CUDA Toolkit 11.8 or 12.1+
# - Matching PyTorch CUDA version
# - C++ compiler
# If flash-attn fails to install, the notebook will fall back to standard attention