Marshall
Marshall
also you might need these bitsandbytes==0.44.0 peft=0.10.0
My shard loading time is very fast, but it seems like there is a 1m delay before it even start loading
I do not get it, what's wrong
I got nan during training, I think it is because I loaded the model as float16?
I found that it's because there is no empty class or floor class in this case. And the resolution needs to be very high to get good results
It is auto included I think from arxiv
It also did not have the leaderboard anymore On Sunday, July 27, 2025, Saunak ***@***.***> wrote: > *Saunak626* left a comment (paperswithcode/paperswithcode-data#45) > > > You can give a try...
I think the data is just on GitHub On Tue, Jul 29, 2025, 16:36 Yeqi Huang ***@***.***> wrote: > *Chivier* left a comment (paperswithcode/paperswithcode-data#45) > > > Does anyone have...