Closed amrrs closed 5 months ago
The example link has been 404. And we don't have a plan to share the shaded model. Therefore, I close the issue.
For anyone else coming here, someone else on HF had already shared a sharded model - https://huggingface.co/notzero/deepseek_sharded/tree/main (I'm not sure of the legitimacy or the content of it, use at your own risk)
Thanks for this. The 7B model can be fit in Google colab given there model are tiny pieces. Example: https://huggingface.co/bn22/Mistral-7B-Instruct-v0.1-sharded