-
-
Hello,
I am trying to run the dalle-flow server container on OpenShift with an Nvidia A40 GPU.
After waiting 10 minutes for dalle and diffusion, the diffusion executer terminates with a timeout.
I …
-
Prior to commit ed91ab4a30ffdcf7e8773e6b434816f79f5fead8
`python image_from_text.py --text="bart simpson on mars" --mega --seed=666 --image_path=bart-simpson-on-mars-666 --torch`
That command w…
-
still digging into it, but something after commit `3e64e86` doubled image generation time again. will update with more info if i can find it
-
Hi,
I can't get the model working using the replicate image (`r8.im/kuprel/min-dalle@sha256:71b9ef81385fae73b632d7e2fe0f5988a739781e833a610a9c83bc45205d8215`) on any GPU because of OOM errors. I tr…
-
docstrings are used in decorated functions (i.e. `@hook.command`) to show help in the chat when people use the command incorrectly. there are a lot of missing docstrings currently, which makes command…
-
Latest branch (using torch, not mega rtx 2080s 8g)
```
It took 0:01:14.897944 to generate F10 images
Average time to generate an image: 0:00:07.489794
```
Rolled back to 3837710
```
It …
-
Hi,
Thanks for this project, you gave me hope to give my AMD GPU some fun :-) (torch is much easier to get to work on it than XLA)
That being said, it doesn't work yet, because the torch variant…
-
Hi!
Was trying to run this project in Ubuntu 20.04.04 (in WSL 2 under Windows 11), and after executing `sh setup.sh` received the following:
```
efim@DESKTOP-PSEQ17Q:/mnt/f/Projects/min-dalle$ sh s…
-
I'm running into unexpected behavior of the text tokenizer, running this on Windows, Python 3.7 , in a virtual environment, using the supplied image_from_text.py script file.
The input text is toke…