Gantavya Bhatt

icon indicating a website link https://sites.google.com/view/gbhatt/

icon company University of Washington icon location Seattle icon location Graduate Student working with Submodularity and NLP.

Results 14 comments of Gantavya Bhatt

Distance Collapse Or Explosion Prevention

Hello, any progress in this issue; or any relevant literature helping to prevent such collapse?

Inconsistent Selector gains and duplicate elements in Selector ranking in case of Facility Locations function in v0.6.0

Hi, Thanks for a quick response. I have a question related to v0.5.0 - Is it OK to use v0.5.0 for the time being? I did not find this issue...

Inconsistent Selector gains and duplicate elements in Selector ranking in case of Facility Locations function in v0.6.0

Alright, I will use v0.5. Are there any other differences between v0.5 and 0.6?

Inconsistent Selector gains and duplicate elements in Selector ranking in case of Facility Locations function in v0.6.0

Thanks for the detailed response!

Incorrect accuracy calculation with DistributedDataParallel - examples/imagenet/main.py

Any update on this? Is this issue still persistent?

DDP: why does every process allocate memory of GPU 0 and how to avoid it?

This still doesn't seem to be helping in my case :-(

Discrepancy between the loss mentioned in the paper and GitHub

Hey, do you mean this in the caption of Figure 4 - "One trajectory (in the center) strongly deviates from typical trajectories seen during training, and the model struggles to...

Strange training dynamics for ImageReward model.

Hello, sorry I couldn't get back with the question on python version 3.10.13 and CUDA 11.7 Experiment was run using torch 1.13.0 Regarding the learning dynamics, I am using the...

Training ImageReward model on different budgets

Thank you very much!

Training ImageReward model on different budgets

Hey will it be possible to provide the hyperparams for 1k and 4k settings as well? That will be very useful.