Brian Qu
In the paper, authors have said that they use sigmoid function to do the classification. They have tried softmax, but sigmoid is better probably.
Actually, you can read the visualize_single_image.py. They will choose the result whose score is larger than 0.5.
Actually, you can put all the scans including both train-val and test to the data-dir because the code will distinguish them automatically. If you process test scans, the code will...
Hello! Thanks for your great work! I also focus on the CLIP convert. In my implement, the activation of OpenAI's CLIP is a little different, using `QuickGELU` instead of `GELU`....
It's been a long time. I remember that I've modify the token_id setting in `generate` function of InstructBLIP. Actually, this doesn't have much impact at all.
Hi, I've tried this before. But the program is stuck. How can I debug this? And I want to know whether it is because I use 30B+ LLM and zero3...
Sorry, it is inconvenient to share the whole code. I would try my best to provide more information. It is a dense model. I've tried the script on my ~9B...
after double check, I find another error message on one worker. as following（time-out error probably）: ``` [E ProcessGroupNCCL.cpp:475] [Rank 15] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=96383, OpType=_ALLGATHER_BASE, NumelIn=88200, NumelOut=5644800, Timeout(ms)=7200000)...
hi, I also test this in one node(8 x A100) with one 9B model. Stuck appeared. TAT
Oh, thanks, I get it. Do you have any suggestion about this? I think I've done left-padding. How to ensure the output length?