s4583222 - Stable Diffusion model (Wrong branch fix) #484
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Model will now train. Have trained over night and so far has a loss of 0.05
This commit is for debugging, getting the code running on Rangpur, as well as a complete reconstruction of my UNET model. It was too big which resulted in the GPU running out of memory. I am trying a different architecture. Also Progress for the README report
I had an issued running my code on gpu's. It would run out of memory after about 3 epochs. I have changed up my code and now my code has been training for 11 hours
I trained for 18 hours and found that it wasn't training correctly. I have gone back a written my training and Unet code
I had an issue with my reverse process. I was not feeding back the noised image
I have added in another time to save the model (but to a separate file for a back up). I let it train overnight and it show how crashed, when I tried loading the model it was corrupted. This is because it would've crashed whilst saving the model. I have created a separate instance of it saving the model. But at a different time to the first save. That way if one becomes corrupted I can load up from a 5 epochs ago. Also have fixed my de noising for my predict file. Which can show images of stuff that look similar to brains.
|
You history looks good, but PR is really late. Can you post link to the original (closed?) PR ASAP? |
shakes76
added
Ungraded
Good Practice (Design/Commenting, TF/Torch Usage)Adequate use and implementation Recognition ProblemSolves problem (some poor recons, but difficult problem) -2 +2 Commit LogMeaningful commit messages could be more descriptive -1 DocumentationReadMe OK, broken links -2 Pull RequestSuccessful Pull Request (Working Algorithm Delivered on Time in Correct Branch) LATE: Waiting for PR link to confirm marks. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Hi all,
My original pull request got deleted when trying to fix the issue of my report being in the wrong branch.
My report was on stable diffusion using the OASIS brain dataset.
My files have been split up into the following
Thanks,
Tom