Fine-tuning instructions

#6
by millionwords - opened

Hi,

It would be great if you could also add fine-tuning instructions and provide the code and sample data for it.

We want to train it to write creative writing and to edit stories and provide feedback

Hello, @millionwords !

There are no official guide for fine-tuning EXAONE for now.

But, you can still use any of popular fine-tuning frameworks; we recommend TRL(https://github.com/huggingface/trl) or native HF trainer for starter.

Also, please aware that our smallest models (LGAI-EXAONE/EXAONE-Deep-2.4B, LGAI-EXAONE/EXAONE-3.5-2.4B-Instruct) use a slightly different architecture with tied embeddings.

Your need to confirm your account before you can post a new comment.

Sign up or log in to comment