Ruohan Gao
Ruohan Gao
Objects before 500 should all be downloadable now. We are doing some re-organization and verification of the rest of the objects and will share them in the coming weeks. Thank...
Hi, We didn't use all 1000 imagenet classes, but ~20 selected audio-related classes. Then we normalize the class probabilities for these classes, so you could get multiple labels with class...
Hi, We use all the collected basis vectors to initialize W, namely M x K with M = 3000, K=25. 3,000 is just a hyperparameter, and a larger number of...
The model trained on split 1 of FAIRPlay is shared at: https://drive.google.com/drive/folders/1qh6PikVtqWXM-pz_2EF6AemoX7Q6tMgQ
Sorry. That link has expired. I lost access to the repo and wasn’t able to update the link and the folder. Here are the models: https://drive.google.com/drive/folders/1fq-SK4OBFVegLM0PfVcLerR9exSwcDWl?usp=sharing