Generally folks have been running stuff on Kaggle, Colab, or Paperspace. For running locally, you’ll need PyTorch, fastai, and the HuggingFace libs installed all with CUDA working.
You might want to check out the docker containers by seeme.ai or paperspace. For Part 1 I used the paperspace container, but I see that the seeme.ai container is newer (pushed a month ago vs paperspace 3 months ago) and smaller than the paperspace container.
Seeme.ai container: Docker Hub
paperspace container: Docker Hub
HTH,
Hi Jeremy,
I have registered for the course, but other than the payment confirmation email, I have not no other information. Can I just view the course here?
Hi Amir
I had a similar confusion. I believe you are all set if you are here and can posts messages into this section of course forum. Check ** About the Part 2 2022 course** for the details on how to get to the live stream.
thanks
Thanks a lot, so the registration was not necessary for me?
Exactly. Browse this link: About the Part 2 2022 course for exact date/time and links for each session.
It was necessary - it’s the reason you have access to this forum category and can watch and participate in the course!
One course content question: Will there be some time devoted to discussing performance optimizations?
It looks like various techniques have been used to reduce the required VRAM for training DreamBooth from 24 GB to under 8 GB! Techniques like these and e.g. gradient accumulation discussed in part one of the course could make the difference in running on consumer hardware.
If time permits would also be very interested in learning more about how to apply sequence models to video (or ct scan slices)… maybe going from resnet activations on individual images/frames to an lstm (or transformer) for overall sequence classification etc in a “graceful” way using fastai (if that’s the best way).
Thanks!
Awesome!
That sure looks like a lot of fun !
For those playing around with the stable_diffusion.ipynb notebook in diffusion-nbs repo running into
RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.cuda.HalfTensor) should be the same
when running pipei2i:
You need a diffusers version higher than 0.4.1, because this patch is needed for fp16 to work.
Why do we try to draw the noise rather than go straight to drawing the digit itself? My thought is that our end goal is to draw the digits in this case, but I’m not quite understanding why we try to draw the noise as our model output rather than just drawing the digit directly
I was wondering whether it would be possible to use this result on zero-shot latent stitching to speed up the CLIP part of the model…
Hi,
I keep getting the below error while running the notebook on colab (free version) with GPU runtime. Any suggestion on how to resolve this?
OSError: There was a specific connection error when trying to load CompVis/stable-diffusion-v1-4:
<class 'requests.exceptions.HTTPError'> (Request ID: o52_DNplzfZM55fVguDXA)
Thanks in advance,
Ganesh
You need to log into a huggingface account and accept the licence terms before you can download stable diffusion (it hasa special licence)
Any specific URL I need to visit? I have logged in and generated a token which was passed to the notebook_login() code.
Thanks for this. Although that alone did not work for me. I had to update Transformers as well.
Click the link for the model in the notebook.