jason law

reading file issue

1

Why I can't read png file from assets? Is that a problem from my module?

Thanks for your excellent work! I want to reproduce the fine-tuned results from flan-v2, I notice that the flan-v2 dataset is not provided in this repo, where can i get...

about finetune_task_lora.sh

8

### Describe the issue What does mm_projector_lr mean in the training script. Does it indicate that the mm_projector will be tuned during lora training? But I don't see the mm_projector...

Grounded VQA

1

Is there any code to support grounded VQA?

Hello, I wonder how to evaluate VQA in the open-ended setting. Can I use the zero-shot evaluation script to evaluate a fine-tuned checkpoint?

About VQA annotations

1

Hello, thanks for your excellent work. I'm reproducing the results in the repo. I found that the vqa_train annotation files differ from the original VQAv2 annotations. There are some answers...

I'm trying to reproduce some results with a downloaded pre-train model.But When I set GPU_ID to 0,1,2,3,the program seems not to run on multiple gpus as I expected.I wonder how...

Can not reproduce results on the table

7

Great thanks for your work! I try exacy the same setting but I got different results on MMLU and BBH. The alpaca-tuned llama always perform worse than original llama(7B or...

Can not reproduce GSM8K zero-shot result

4

Thanks for your work! I try to reproduce the GSM8K results within this project. I simply removed the code of tranferring models to peft to achieve this. However, I can't...

Problems during run initial step

12

Here is my running code: import evaluate ev = evaluate.load('exact_match') Always return below error Downloading builder script: 3.69kB [00:00, 4.56MB/s] Traceback (most recent call last): File "", line 1, in...