TreeForest

icon company University of Illinois Chicago icon location Chicago icon location CS PhD Student @ UIC | Applied Scientist Intern @ AWS AI 24' @ Amazon 23'22' | LLMs&NLP&CV

Results 5 comments of TreeForest

training log

I have the same question here

Further Explanation of Eqn(5)

Could someone help to give some insights or explanation on it?

Issue with "kv_cache" while using modified generate/lora.py for a list of inputs

When conducting generation for multiple consecutive inputs on a LoRA fine-tuned LLaMA, I noticed that using 'reset_cache' after each generation for one input will affect the quality of generation on...

4-bit multi-gpu training

Hi Yingyue, have you figured out how to properly do 4-bit multi-gpu training? Can you obtain similar results with the number reported in the paper?

Author Page: Henry Peng Zou

Hi @nschneid, thanks for your quick response and guidance. I have submitted request to fix metadata for those paper indidually: 1. https://github.com/acl-org/acl-anthology/issues/5383 2. https://github.com/acl-org/acl-anthology/issues/5384 3. https://github.com/acl-org/acl-anthology/issues/5385 4. https://github.com/acl-org/acl-anthology/issues/5386