Nothing Happens after clicking generate button (5700XT/Linux)

[hereforsd5700]

GPU: RX 5700 XT
OS 1: Ubuntu 23.04 (Kubuntu/KDE flavour)
OS 2: Windows 10 LTSC
Python: 3.11.2 (comes bundled with Ubuntu)

My system is a dual-boot system and I have gotten lshqqytiger's DirectML fork working on Windows 10. So I decided to give a try to ROCm.

What I have done so far:
With multiple troubleshooting attempts, I managed to get the UI working in Ubuntu. Which involved:

• Modified webui-user.sh, and added PATH for TCMalloc:

PATH=$PATH:/usr/sbin
export COMMANDLINE_ARGS="--precision full --no-half --medvram"
export TORCH_COMMAND="pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm5.4.2"

• Installed drivers from AMD repo: amdgpu-install_5.3.50300-1_all.deb

• Gave chmod 777 permission to webui.sh and webui-user.sh

• Gave myself permissions to render and video
  sudo usermod -a -G render
  sudo usermod -a -G video

• Arguments: --precision full --no-half --medvram
  The models currently in Stable-diffusion directory are: v1-5-pruned-emaonly.safetensors and reliberate_v10.safetensors

What happens:

When I execute webui.sh, console shows this:

################################################################
Install script for stable-diffusion + Web UI
Tested on Debian 11 (Bullseye)
################################################################

################################################################
Running on <Username>
################################################################

################################################################
Repo already cloned, using it as install directory
################################################################

################################################################
Create and activate python venv
################################################################

################################################################
Launching launch.py...
################################################################
Using TCMalloc: libtcmalloc.so.4
Python 3.11.2 (main, May 30 2023, 17:45:26) [GCC 12.2.0]
Version: v1.3.2
Commit hash: baf6946e06249c5af9851c60171692c44ef633e0
Installing requirements
Launching Web UI with arguments: --precision full --no-half --medvram
No module 'xformers'. Proceeding without it.
Loading weights [6ce0161689] from /home/<Username>/stable-diffusion-webui/models/Stable-diffusion/v1-5-pruned-emaonly.safetensors
Running on local URL:  http://127.0.0.1:7860

To create a public link, set `share=True` in `launch()`.
Startup time: 3.0s (import torch: 0.6s, import gradio: 0.6s, import ldm: 0.7s, other imports: 0.4s, load scripts: 0.2s, create ui: 0.2s, gradio launch: 0.2s).
Creating model from config: /home/<Username>/stable-diffusion-webui/configs/v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
Applying optimization: Doggettx... done.
Textual inversion embeddings loaded(0): 
Model loaded in 1.4s (load weights from disk: 0.5s, create model: 0.3s, apply weights to model: 0.4s, load VAE: 0.2s).

After this I open the UI in Firefox, gave a short prompt, and hit generate. Nothing happens. I've waited for 10-15 minutes, but nothing in the UI or the console. No progress bars or anything.

However my GPU usage goes up so something is happening, but I don't know what.
(Screenshot:
)

After waiting for 10-15 minutes I terminate the sh and it gives me this:

^CInterrupted with signal 2 in <frame at 0x7efe43d651c0, file '/usr/lib/python3.11/threading.py', line 324, code wait>

And GPU usage drops to normal.

In case you're wondering, the screenshots are from a monitoring utility called RandR

Any help would be great. Thank you.

[49RpK5dY]

Pytorch compiled with rocm5.4.2 won't work on 5700. You need TORCH_COMMAND="pip install torch==1.13.0+rocm5.2 torchvision==0.14.0+rocm5.2 --extra-index-url https://download.pytorch.org/whl/rocm5.2"

There is also a way to get torch2.1 working with rx5700, this should work with a1111 if you download torch and torchvision and point TORCH_COMMAND to local files.