search

unslothai / unsloth

Finetune Llama 3.2, Mistral, Phi & Gemma LLMs 2-5x faster with 80% less memory
https://unsloth.ai
Apache License 2.0
16.28k stars 1.12k forks source link

Confusion about the ModelFile #1075

Open Ambitious4 opened 2 days ago

Ambitious4 commented 2 days ago

Hi, I want to upload finetune Llama 3-instruct to ollama, and following the https://docs.unsloth.ai/tutorials/how-to-finetune-llama-3-and-export-to-ollama to do it, but it didn't generate the ModelFile.

And I see > #798 , then I use the apply_chat_template to do, but I still can't get the ModelFile. :(

Hope to get your help.

danielhanchen commented 1 day ago

@Ambitious4 Apologies on the delay!

We have 2 notebooks for Ollama - which might be helpful:

  1. General Ollama exporting: https://colab.research.google.com/drive/1WZDi7APtQ9VsvOrQSSC5DDtxq159j8iZ?usp=sharing
  2. CSV / Excel files to Ollama: https://colab.research.google.com/drive/1VYkncZMfGFkeCEgN2IzbZIKEDkyQuJAS?usp=sharing