Seungyoun, Shin

you are right. I have big mistake. I will update it as soon as possible.

I dont know why this error occured. It's probably different torch version creating zero_tensor in different way. I will check it as soon as possible

This is the result from "COCO 2014 train" , almost the same dalle structure @ 1epoch I notice that after 1epoch loss doesn't seem to decrease and the image is...

Do we have train code for hf-owl-vit?

batch size 1 and gradient accumulation 1 works for more than 2000 steps. Not sure I can finish training without OOM

I also posted https://github.com/tatsu-lab/stanford_alpaca/files/11024692/Mar20_05-17-08_0c56f6779a08.csv csv log file! I takes approx. 24hours(a day) This is strange. You are using way better gpus than mine. As you mentioned wandb could be the...

--bf16 True \ --per_device_train_batch_size 1 \ --per_device_eval_batch_size 1 \ --gradient_accumulation_steps 1 \ @danwei1992 Make sure this 4 terms are correct I am using - docker image :[hf:latest]( https://hub.docker.com/r/huggingface/transformers-pytorch-gpu) - CUDA...

This is my experiments with prompt tunning. Not perfect but pretty amazing Seems like img1,img2,text is better performing.

 Deleting the line and build from source works. I am not sure this is desired or not + At ProcThor, we cannot goto some locations even if the location...

\I have successfully modified the Unity code in Fill.cs to disable the spilling action. However, I am still uncertain about why the liquid spills when teleporting without tilting the held...