How to train a layer before the first layer of existing models (e.g. gemma) using axolotl? #2733
Answered
by
NanoCode012
GeorgeCarpenter
asked this question in
Q&A
Replies: 2 comments 5 replies
-
Hey, do you know the layer path? We support freezing the model for FFT and unfreezing certain params via regex: axolotl/examples/mistral/bigstral-ds-zero3.yaml Lines 10 to 16 in 4a8af60 |
Beta Was this translation helpful? Give feedback.
0 replies
-
What is |
Beta Was this translation helpful? Give feedback.
5 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
In our cases, we wanna train a layer before the first layer of existing LLM models (e.g. gemma) using axolotl. The new added layer is used to process the image. What can we do to implement this?
Beta Was this translation helpful? Give feedback.
All reactions