Do pytorch have somewhere leaderboard/benchmark for testing minor modifications?

Would be nice to have premade env, data, and model so I can just test if my modification makes sense.
Of course, kaggle comes to mind, but I want to test GAN model, besides the Kaggle competition usually does not have enough data and in my experience, the influence of random seed is great (and ofc things like fine-tuning to test data, insane ensembling, etc).