[Proposal] Reduce training time by resampling beforehand

[Craq]

The biggest bottleneck in your model, regarding speed, is the Resample layer in the Hierarchical Discriminator. The speed on it is just awful. If you resample your dataset beforehand and load it from disk, your training will be up to 3 times faster.

[rishikksh20]

I tried it but still not able to get good speed on baseline model.