Nick Konz

@bsamseth how did you edit your Makefile specifically to get this to work? I'm still having the same issue on macOS Catalina.

It might be an issue with this code needing an older `pytorch_lightning` version. Try installing version 1.6.1 (this may create other problems but should solve this.

Unless you want to compute the shortest path MLE, you can get around this by simply commenting out `from .shortest_path import shortest_path` in line 4 of `estimators/__init__.py`.

Hi, sorry to hear that this issue is happening. How many epochs did you train for, and what do the images sampled during training look like in the `samples` folder...

Hi, I found a bug where the number of network output channels was not set properly; fixed in commit 28592394398dd2590c72b96aa0bad9b8b1d36f91. I tested the fix by training an unconditional model from...

Hi, My apologies for the late reply, this fell through the cracks! For your first problem, I haven't seen this issue and have had no issues doing multi-GPU training myself....

Hi, I'm happy to see that you're trying out our model on cell images! I think that this bug may be due to the code being designed for one-channel images....

Hi, I modified the code to try to load your images from np.arrays, rather than through PIL, if the chosen channel count in the `--num_img_channels` argument of `main.py` is not...

Thanks for debugging this! Your solution is good, I hadn't considered those errors; I'll go ahead and add your fixes as a commit 9f532bb98e328301d48218d3257e8db936e59cb3 (or let me know if you...

Hi, My apologies for the late reply! The setting `class_conditional` at https://github.com/mazurowski-lab/segmentation-guided-diffusion/blob/6ae1b3e832f53e09985d37efc2fa372f8ba38516/training.py#L50 is experimental (not fully tested), and was only for testing **class**-conditional classifier-free guided image generation. In your case...