Is it possible to customize an MOE model configuration in the ablations directory and train it?
- Dominant language
- Jupyter Notebook
- Stars
- 1.1k
- Forks
- 120
- PR merge metrics
- No merged PRs in 30d
Description
Hello,
I am currently working with this repository and I am interested in customizing a mixture-of-experts (MOE) model configuration within the `ablations `directory. Specifically, I would like to modify certain aspects of the model and training parameters to create a customized MOE model for training.
Could you please confirm if it is possible to create and train a custom MOE configuration in the `ablations `directory? If so, I would appreciate any guidance or suggestions on how to modify the configuration and proceed with the training process.
Thank you in advance for your help!
Best regards,
Rongrong Zhou
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.