pace577

The following seems to run, but is it recommended? ``` python set_trainable(model, recompile=False) model.compile( model.optimizer, loss=model.loss, metrics=model.metrics, loss_weights=None, sample_weight_mode=None, weighted_metrics=None, ) ```

@wow-yorick I had the same problem and I'm using Doom Emacs. How did you solve it? I added the following line to the `:config` of my `use-package` block (saw that...

Changing `self.discrete_sigmas[timestep - 1].to(t.device)` to `self.discrete_sigmas.to(t.device)[timestep - 1]` in [this line of sde_lib.py](https://github.com/yang-song/score_sde_pytorch/blob/cb1f359f4aadf0ff9a5e122fe8fffc9451fd6e44/sde_lib.py#L251) seems to fix the problem.