Starting with the models published as Open Source (not published the details about the training):
Reproducing the DeepSeek-R1 creation process from scratch: https://github.com/huggingface/open-r1
Next steps:
- Understand the relation of DeepSeek-V3 with other models
- Play locally with DeepSeek-V3 and understand limitations of inference locally and how to scale it
- Understand why DeepSeek-R1 based on DeepSeek-V3 is so disruptive
Starting with the models published as Open Source (not published the details about the training):
Reproducing the DeepSeek-R1 creation process from scratch: https://github.com/huggingface/open-r1
Next steps: