Image-Text-to-Text
Safetensors
multilingual
deepseek_vl_v2
deepseek
unsloth
vision-language
ocr
custom_code
Instructions to use unsloth/DeepSeek-OCR with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Local Apps
- Unsloth Studio new
How to use unsloth/DeepSeek-OCR with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for unsloth/DeepSeek-OCR to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for unsloth/DeepSeek-OCR to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for unsloth/DeepSeek-OCR to start chatting
Load model with FastModel
pip install unsloth from unsloth import FastModel model, tokenizer = FastModel.from_pretrained( model_name="unsloth/DeepSeek-OCR", max_seq_length=2048, )
DeepSeek-OCR Fine-tuning is now available!
#2
pinned
by danielhanchen - opened
Hey guys, you can now fine-tune DeepSeek-OCR locally or for free with our Unsloth notebook. Unsloth GitHub: https://github.com/unslothai/unsloth
- For the notebook, we showcased how fine-tuning DeepSeek-OCR with a Persian dataset, improved its language understanding by 88.64%, and reduced Character Error Rate (CER) from 149% to 60%.
- The 88.64% improvement came from just 60 training steps (if you train longer it'll be even better). Evaluation results in our blog.
- ⭐ If you'd like to learn how to Run/fine-tune DeepSeek-OCR or know details on the evaluation results etc., you can read our guide here: https://docs.unsloth.ai/new/deepseek-ocr
- DeepSeek-OCR free Fine-tuning notebook: https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Deepseek_OCR_(3B).ipynb
Thank you so much and let me know if you have any questions! :)
danielhanchen pinned discussion