Instructions to use JeffersonNunn/deepseek-ocr-metal with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use JeffersonNunn/deepseek-ocr-metal with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("JeffersonNunn/deepseek-ocr-metal", dtype="auto") - Notebooks
- Google Colab
- Kaggle
How to run this?
#2
by jondecker76 - opened
Can you show some example code to run this on Apple Silicon?
It was my first time trying to do anything like this. I messed up the push and my laptop is not here lol …
I will post tomorrow a complete copy of what I did. Yesterday I was working on an MLX version
Is it mostly the same as the transformers code on the main deepseek ocr page?
Well. IDK how I managed to run it once but not since. I keep getting errors.
I spent quite a bit of time yesterday trying to getting it running…
I had a patch that basically wrapped it and ran it through metal
I’ll have to try it again next weekend…