TextRL
TextRL copied to clipboard
token classification test
Tried to create notebook in examples folder for token classification problem. Please help me develop this.
Can you explain how reinforcement learning can be used for token classification? I would appreciate more information on this topic.
I tried doing it but failed. But you can find here the failed attempt and have some leads. Actually I want to have RLHF for as many tasks as possible from only token classification to image classification, image segmentation, etc etc
If you want to collaborate we can work together on this and make a adaptor type model that adds RLHF extension on the output side of any model and instead of human labbeled data it does RLHF for supervised learning