Naihao Deng
Naihao Deng
Yeah, some of the packages are out of data (e.g. imresize) and you need to modify the code a little bit to get it to work. And the code might...
No, sorry. It is a bit weird to me since I am pretty sure every place I changed is consistent with prior works.
I got the same error, any update on this issue?
Thanks a lot for sharing @LeSphax! I am still in the middle of debugging myself, and my output is similar to yours, I got it at the start of my...
Any update on this issue? I am using Pytorch Lightning, originally I thought it is because I am passing too many things for each step, but after I change those,...
Thanks @teaguexiao, I will try using more GPUs (but ours are A40 of 48 GB memory each) to see if that can help. Thanks for sharing!
> It seems like using capital letters with the model name `7b` or `7B` makes the difference. > > It worked with capital letters for me and gave 403 for...
@mertunsall @YunfanZhang42 Not sure if this is the right place to ask. Have you observed any discrepancy between the validation outputs versus the generation outputs? For training I have ```...
@YunfanZhang42 I see, thanks for the clarification! For the RL, yes I understand it would need temperature != 0 for exploration. I think `actor_rollout_ref.rollout.val_kwargs.temperature=0` would only affect the validation setup...
Oh gotcha, that is for the `main_generation.py` and I was trying to align the results from the generation with the results from validation step = 0 during training. But it...