Added images and details for PPO training

HumanSignal / RLHF

Collection of links, tutorials and best practices of how to collect the data and build end-to-end RLHF system to finetune Generative AI models

175 stars 36 forks source link

Closed JimmyWhitaker closed 1 year ago

JimmyWhitaker commented 1 year ago

Simplified notebook name
Added illustrations to notebook
Added details on PPO training for the notebook (Note: trlX limitations prevent training this model in Colab)

niklub commented 1 year ago

Could we replace the image by dragging the left sidebar to the left, to align 2 options horizontally instead of vertically?

niklub commented 1 year ago

Is it still WIP or I can remove it?