jocastrocUnal

Results 8 comments of jocastrocUnal

[Bug] Fail to install (with python 3.10.13)

Same problem here

Cannot merge LORA layers when the model is loaded in 8-bit mode

I hope in the future this code could work... its more natural. ```python model = peft_model.merge_and_unload() model.save_pretrained("/model/trained") ```

Is there an API available for this program?

I like that to :)

Bug on Ubuntu 22: software doesnt work

same here ![image](https://github.com/PromtEngineer/localGPT/assets/93233286/63ab95b5-960d-4382-8517-2b747c78a65f)

Loss drop in cero

> I also have the same problem. The loss first decreases, then it slowly grows till it drops to zero. The image below shows the training loss in the first...

Evaluation loss issue (strange trend)

same here ![image](https://github.com/artidoro/qlora/assets/93233286/57835276-e127-4e86-9421-dc801de57361)

how to continue training?

In the training arguments I set `resume_from_checkpoint = True`. But this is for huggingface. https://huggingface.co/docs/transformers/main_classes/trainer#transformers.TrainingArguments

Cache only has 0 layers, attempted to access layer with index 0

The same issue here. With the model "llama-3-8b-Instruct-bnb-4bit" in here. > I also just ran into this exact same issue. The model I am using is > > https://huggingface.co/NousResearch/Hermes-2-Pro-Llama-3-8B >...