f12f0200d6
parametric efficiency improvements and lets try feeding the timestep into the conditioning encoder |
||
|---|---|---|
| .. | ||
| audio | ||
| classifiers | ||
| clip | ||
| composable | ||
| diffusion | ||
| image_generation | ||
| image_latents | ||
| lucidrains | ||
| optical_flow | ||
| vqvae | ||
| __init__.py | ||
| arch_util.py | ||