stable-diffusion-rocm-docker
stable-diffusion-rocm-docker copied to clipboard
l1na-forever
CodeFormer/PyTorch issue
Docker on Ubuntu 22LTS - AMD 6900XT GPU
Fresh pull today, fails to start, complaining about setting up CodeFormer. tore down, re-pulled from scratch, same error:
Error setting up CodeFormer:
Traceback (most recent call last):
File "/sd/modules/codeformer_model.py", line 38, in setup_model
from facelib.utils.face_restoration_helper import FaceRestoreHelper
File "/sd/repositories/CodeFormer/facelib/utils/face_restoration_helper.py", line 7, in <module>
from facelib.detection import init_detection_model
File "/sd/repositories/CodeFormer/facelib/detection/__init__.py", line 11, in <module>
from .yolov5face.face_detector import YoloDetector
File "/sd/repositories/CodeFormer/facelib/detection/yolov5face/face_detector.py", line 20, in <module>
IS_HIGH_VERSION = tuple(map(int, torch.__version__.split('+')[0].split('.'))) >= (1, 9, 0)
ValueError: invalid literal for int() with base 10: '0a0'
WARNING:root:Pytorch pre-release version 1.13.0a0+gitd566718 - assuming intent to test it
Traceback (most recent call last):
File "launch.py", line 169, in <module>
start_webui()
File "launch.py", line 164, in start_webui
webui.webui()
File "/sd/webui.py", line 92, in webui
initialize()
File "/sd/webui.py", line 85, in initialize
shared.sd_model = modules.sd_models.load_model()
File "/sd/modules/sd_models.py", line 181, in load_model
sd_model = instantiate_from_config(sd_config.model)
File "/sd/repositories/stable-diffusion/ldm/util.py", line 85, in instantiate_from_config
return get_obj_from_str(config["target"])(**config.get("params", dict()))
File "/sd/repositories/stable-diffusion/ldm/util.py", line 93, in get_obj_from_str
return getattr(importlib.import_module(module, package=None), cls)
File "/opt/conda/lib/python3.7/importlib/__init__.py", line 127, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "<frozen importlib._bootstrap>", line 1006, in _gcd_import
File "<frozen importlib._bootstrap>", line 983, in _find_and_load
File "<frozen importlib._bootstrap>", line 967, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 677, in _load_unlocked
File "<frozen importlib._bootstrap_external>", line 728, in exec_module
File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
File "/sd/repositories/stable-diffusion/ldm/models/diffusion/ddpm.py", line 19, in <module>
from pytorch_lightning.utilities.distributed import rank_zero_only
ImportError: cannot import name 'rank_zero_only' from 'pytorch_lightning.utilities.distributed' (/opt/conda/lib/python3.7/site-packages/pytorch_lightning/utilities/distributed.py)
Just to confirm - fresh Ubuntu 22 install on empty drive, installed docker, and pulled [stable-diffusion-rocm-docker] - errors out the same way. Something new is broken in the last 2 weeks (since I set this up the last time)
To add, I am also experiencing this error on Manjaro. I haven't had this git working as of yet as I have only tried to install it today.
I have managed to get past this point with a standard installation before.
Edit:
Interestingly I have managed to get it to launch using the command:
drun --name sd l1naforever/stable-diffusion-rocm:baked
Though it still carries the same error, and wont actually produce anything, even though the webui is working
Unable to find image 'l1naforever/stable-diffusion-rocm:baked' locally
baked: Pulling from l1naforever/stable-diffusion-rocm
3b65ec22a9e9: Already exists
672c6966ed08: Already exists
c6ff2848b7a2: Already exists
16b13c05c444: Already exists
c6eddca520ee: Already exists
f1b9a9cad1c7: Already exists
aa964431ec1f: Already exists
19f05ac8f4fd: Already exists
1bb04df460f1: Already exists
a00eb1dc88fa: Already exists
c65f0f6e3621: Already exists
4a0073e28e75: Already exists
51db81038349: Already exists
8b30faa6eea7: Already exists
ec9a5f407a44: Already exists
b74e875ede84: Already exists
62dc6a9ee118: Already exists
dc832aafeaf6: Already exists
e279ebfd5e92: Already exists
f8802f99fcf9: Already exists
5dc3d6baee95: Already exists
0a74aad6f260: Already exists
18ac3259995c: Already exists
329ba27edb65: Already exists
ff4708d15eb2: Already exists
65e97f4feb3a: Already exists
5044019df697: Already exists
fab2692dec7a: Already exists
2996f6ef999f: Already exists
a6ec0f016a81: Already exists
50a6bf13e1d5: Already exists
9c40611e59c9: Already exists
fedc0c0800e1: Already exists
02441ba6f3a1: Already exists
d18712e2f093: Already exists
ffe880ee0557: Already exists
0cf564656846: Already exists
10b6cdb90d0f: Pull complete
Digest: sha256:2ecd15a4e445cc95d511853d387ddb5ca4c4ae09c2a20643e8ab0f827cec71fe
Status: Downloaded newer image for l1naforever/stable-diffusion-rocm:baked
Python 3.7.13 (default, Mar 29 2022, 02:18:16)
[GCC 7.5.0]
Commit hash: 08b3f7aef15f74f4d2254b1274dd66fcc7940348
Installing requirements for Web UI
Launching Web UI with arguments: --precision full --no-half
Error setting up CodeFormer:
Traceback (most recent call last):
File "/sd/modules/codeformer_model.py", line 38, in setup_model
from facelib.utils.face_restoration_helper import FaceRestoreHelper
File "/sd/repositories/CodeFormer/facelib/utils/face_restoration_helper.py", line 7, in
WARNING:root:Pytorch pre-release version 1.13.0a0+gitd566718 - assuming intent to test it LatentDiffusion: Running in eps-prediction mode DiffusionWrapper has 859.52 M params. making attention of type 'vanilla' with 512 in_channels Working with z of shape (1, 4, 32, 32) = 4096 dimensions. making attention of type 'vanilla' with 512 in_channels Loading weights [7460a6fa] from /sd/models/Stable-diffusion/model.ckpt Global Step: 470000 Applying cross attention optimization (Doggettx). Model loaded. Loaded a total of 0 textual inversion embeddings. Running on local URL: http://127.0.0.1:7860
To create a public link, set share=True in launch().`
The specific error was fixed in https://github.com/sczhou/CodeFormer/commit/8e5f81be41f283942b4346fef87ca1fad2cb889d, but I couldn't find the reason why this container still pulls the old version
