Taha Bagheri

icon location London, UK icon location MSc AI @ Queen Mary University of London

Results 3 comments of Taha Bagheri

Would you kindly update Xlora to support Quantized Models?

Hi @EricLBuehler , Just wanted to make sure that the current version supports Quantised models since I think some tests haven't been passed [here](https://github.com/EricLBuehler/xlora/pull/25), and the commit hasn't been merged...

`PPOv2Trainer` and `RLOOTrainer`: Remove the implicit assumption that the reward model & policy model share the same tokenizer

If someone struggles with this, it can be fixed easily. All you have to do is override the get_reward() function: First, decode the generated input_ids using the tokenizer of the...

`PPOv2Trainer` and `RLOOTrainer`: Remove the implicit assumption that the reward model & policy model share the same tokenizer

Hey there, I can fix this and submit a PR if you guys need it . 🧐