Spaces:
Configuration error
Configuration error
| # Copyright (c) 2023 Amphion. | |
| # | |
| # This source code is licensed under the MIT license found in the | |
| # LICENSE file in the root directory of this source tree. | |
| import torch.nn as nn | |
| from abc import abstractmethod | |
| class UNetBlock(nn.Module): | |
| r"""Any module where forward() takes timestep embeddings as a second argument.""" | |
| def forward(self, x, emb): | |
| r"""Apply the module to `x` given `emb` timestep embeddings.""" | |