Can you attach code repository
#1
by
bigman15
- opened
Can you attach code repository on how you fine tuned gpt-j-6b on Shakespeare dataset
I used this code, https://colab.research.google.com/drive/1ft6wQU0BhqG5PRlwgaZJv2VukKKjU4Es (from https://huggingface.co/hivemind/gpt-j-6B-8bit), mentioned in the training procedure section of this model card, but would now recommend checking out PEFT for parameter efficient finetuning methods (https://github.com/huggingface/peft), as this code is out of date
crumb
changed discussion status to
closed