Avinab Saha

icon indicating a website link https://avinabsaha.github.io/ icon indicating an email [email protected]

icon company PhD Student, UT Austin icon location Austin, TX icon location Self-Supervised Learning, Visual Perception, Generative AI

Results 15 comments of Avinab Saha

Pretrained Models

@yustiks please let us know about the performance of your trained model in comparison to the results mentioned in the paper.

Fix Dependency Name in requirements.txt

Try using --head mlp!

could you provide an example of test code with a trained model?

Please use demo_content_aware_feats.py and demo_content_quality_aware_feats.py and merge the features to obtain final ReIQA features (8192 length) for any image! Typically, for IQA/VQA tasks, we develop feature extractors that can be...

Information for training regressors

I will try to have something on this soon. Using sklearn's elasticNet with moderate hyper-parameter will be able to produce results reported in paper. For exact results, need extensive hyper...

Open-Sora 2.0 exceeds 60GB ram usage

@SimonWXW, what is a recommended GPU type for single-GPU inference? I am running out of GPU VRAM when trying to generate the 768px video with 129 frames with an H100.