stable-diffusion-webui
stable-diffusion-webui copied to clipboard
Published
20 hours ago •
AUTOMATIC1111
AUTOMATIC1111
[Bug]: RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.cuda.HalfTensor) should be the same (RTX 2080Ti)
Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
What happened?
Generate process crash upon finishing.
Steps to reproduce the problem
-
git pull
-
pip install -r requirements.txt
-
Files installed manually:
user@linuxuser:~/stable-diffusion-webui$ find ./models -name 768*
./models/Stable-diffusion/768-v-ema.yaml
./models/Stable-diffusion/768-v-ema.ckpt
-
launch './webui.sh' without arguments
-
Action taken: Enter prompt, set options and hit 'Generate'.
-
Error after generation bar fulfilled.
What should have happened?
A photo of a dog should appeared in ./output/txt2img-images/
Commit where the problem happens
Commit hash: 9e78d2c419732711e984c4478af15ece121d64fd
What platforms do you use to access UI ?
Linux
What browsers do you use to access the UI ?
Mozilla Firefox
Command Line Arguments
No arguments on command line.
Additional information, context and logs
OS version: Ubuntu 22.04.1 LTS
Output log:
linuxuser@computer:~/stable-diffusion-webui$ ./webui.sh
################################################################
Install script for stable-diffusion + Web UI
Tested on Debian 11 (Bullseye)
################################################################
################################################################
Running on linuxuser user
################################################################
################################################################
Repo already cloned, using it as install directory
################################################################
################################################################
Create and activate python venv
################################################################
################################################################
Launching launch.py...
################################################################
Python 3.8.13 (default, Mar 28 2022, 11:38:47)
[GCC 7.5.0]
Commit hash: 6074175faa751dde933aa8e15cd687ca4e4b4a23
Installing requirements for Web UI
Installing requirements for Dreambooth
Checking Dreambooth requirements.
Dreambooth revision is c589a3596ade64228de8a7851f50c2470c7a76aa
Args: ['extensions/sd_dreambooth_extension/install.py']
[*] Diffusers version is 0.7.2.
[*] Torch version is 1.12.1+cu113.
[*] Torch vision version is 0.13.1+cu113.
[*] Transformers version is 4.21.0.
[*] Xformers
Launching Web UI with arguments:
WARNING:root:A matching Triton is not available, some optimizations will not be enabled.
Error caught was: No module named 'triton'
Patching transformers to fix kwargs errors.
Dreambooth API layer loaded
Loading config from: /home/linuxuser/stable-diffusion-webui/models/Stable-diffusion/768-v-ema.yaml
LatentDiffusion: Running in v-prediction mode
DiffusionWrapper has 865.91 M params.
Loading weights [2c02b20a] from /home/linuxuser/stable-diffusion-webui/models/Stable-diffusion/768-v-ema.ckpt
Global Step: 140000
Applying cross attention optimization (Doggettx).
Model loaded.
Loaded a total of 0 textual inversion embeddings.
Embeddings:
Running on local URL: http://127.0.0.1:7860
To create a public link, set `share=True` in `launch()`.
100%|███████████████████████████████████████████| 20/20 [00:06<00:00, 3.04it/s]
Error completing request████████████████████████| 20/20 [00:04<00:00, 4.62it/s]
Arguments: ('A photo of a dog', '', 'None', 'None', 20, 0, False, False, 1, 1, 7, -1.0, -1.0, 0, 0, 0, False, 768, 768, False, 0.7, 0, 0, 0, False, False, False, '', 1, '', 0, '', True, False, False) {}
Traceback (most recent call last):
File "/home/linuxuser/stable-diffusion-webui/modules/ui.py", line 169, in f
res = list(func(*args, **kwargs))
File "/home/linuxuser/stable-diffusion-webui/webui.py", line 58, in f
res = func(*args, **kwargs)
File "/home/linuxuser/stable-diffusion-webui/modules/txt2img.py", line 49, in txt2img
processed = process_images(p)
File "/home/linuxuser/stable-diffusion-webui/modules/processing.py", line 430, in process_images
res = process_images_inner(p)
File "/home/linuxuser/stable-diffusion-webui/modules/processing.py", line 534, in process_images_inner
x_samples_ddim = decode_first_stage(p.sd_model, samples_ddim)
File "/home/linuxuser/stable-diffusion-webui/modules/processing.py", line 368, in decode_first_stage
x = model.decode_first_stage(x)
File "/home/linuxuser/stable-diffusion-webui/venv/lib/python3.8/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context
return func(*args, **kwargs)
File "/home/linuxuser/stable-diffusion-webui/repositories/stable-diffusion-stability-ai/ldm/models/diffusion/ddpm.py", line 826, in decode_first_stage
return self.first_stage_model.decode(z)
File "/home/linuxuser/stable-diffusion-webui/repositories/stable-diffusion-stability-ai/ldm/models/autoencoder.py", line 90, in decode
dec = self.decoder(z)
File "/home/linuxuser/stable-diffusion-webui/venv/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/home/linuxuser/stable-diffusion-webui/repositories/stable-diffusion-stability-ai/ldm/modules/diffusionmodules/model.py", line 631, in forward
h = self.mid.attn_1(h)
File "/home/linuxuser/stable-diffusion-webui/venv/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/home/linuxuser/stable-diffusion-webui/repositories/stable-diffusion-stability-ai/ldm/modules/diffusionmodules/model.py", line 267, in forward
out = self.proj_out(out)
File "/home/linuxuser/stable-diffusion-webui/venv/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/home/linuxuser/stable-diffusion-webui/venv/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 457, in forward
return self._conv_forward(input, self.weight, self.bias)
File "/home/linuxuser/stable-diffusion-webui/venv/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 453, in _conv_forward
return F.conv2d(input, weight, bias, self.stride,
RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.cuda.HalfTensor) should be the same
I think same here i've already post my bug in another post i hope here it is the good one for my bug Error completing request Arguments: ('a dog', '', 'None', 'None', 20, 0, False, False, 1, 1, 7, -1.0, -1.0, 0, 0, 0, False, 768, 768, False, 0.7, 0, 0, 0, 0.9, 5, '0.0001', False, 'None', '', 0.1, False, 0.9, 5, '0.0001', False, 'None', '', 0.1, False, '
Running in txt2img mode:
Render these video formats:
Animation Parameters
', 'Initial Parameters
', 'Prompt Template, applied to each keyframe below
', 'Props, Stamps
', 'Supported Keyframes:
time_s | source | video, images, img2img | path
time_s | prompt | positive_prompts | negative_prompts
time_s | template | positive_prompts | negative_prompts
time_s | prompt_from_png | file_path
time_s | transform | zoom | x_shift | y_shift | rotation
time_s | seed | new_seed_int
time_s | noise | added_noise_strength
time_s | denoise | denoise_value
time_s | set_text | textblock_name | text_prompt | x | y | w | h | fore_color | back_color | font_name
time_s | clear_text | textblock_name
time_s | prop | prop_name | prop_filename | x pos | y pos | scale | rotation
time_s | set_stamp | stamp_name | stamp_filename | x pos | y pos | scale | rotation
time_s | clear_stamp | stamp_name
time_s | col_set
time_s | col_clear
time_s | model | 768-v-ema, Anything-V3.0-pruned, Cyberpunk-Anime-Diffusion, Mdjrny-pprct_step_7000, arcane-diffusion-v3, arcane-v4 le-mien, classicAnim-v1, comic-diffusion, comic-diffusion-V2, discoElysium-pruned, discodiffusion, eldenRing-v3-pruned, eldenring-v2-pruned, f222, ghibli-diffusion-v1, mdjrny-v4, model, nitroDiffusion-v1, redshift-diffusion-v1, robo-diffusion-v1, sd-v1-5-inpainting, spiderverse-v1-pruned, superhero-diffusion, v1-5-pruned-emaonly, wulf1
Still happening with hash 0b5dcb3d7ce397ad38312dbfc70febe7bb42dcc3
I wondering if reinstalling may solve the problem; did anyone try it?
Having the same fatal failure with a RTX 3090 on Ubuntu Linux. A fresh clean cloned SD directory does NOT solve the failure. the model .ckpt does not matter.
git bisect shows the regression happened when SD 2.0 was added.
Bisecting: 0 revisions left to test after this (roughly 0 steps)
[ce6911158b5b2f9cf79b405a1f368f875492044d] Add support Stable Diffusion 2.0
Still not working on hash 44c46f0ed395967cd3830dd481a2db759fda5b3b
I was able to fix this by forcing the devices.autocast function to run via a one-line change in modules/processing.py (if you're not familiar with git diffs, see the lines prefixed with - and + for the code change):
diff --git a/modules/processing.py b/modules/processing.py
index 24c537d..efdfcb9 100644
--- a/modules/processing.py
+++ b/modules/processing.py
@@ -396,7 +396,7 @@ def create_random_tensors(shape, seeds, subseeds=None, subseed_strength=0.0, see
def decode_first_stage(model, x):
- with devices.autocast(disable=x.dtype == devices.dtype_vae):
+ with devices.autocast():
x = model.decode_first_stage(x)
return x
My environment:
[ec2-user@ip-172-31-1-148 stable-diffusion-webui]$ python3.8 -m torch.utils.collect_env
Collecting environment information...
PyTorch version: 1.12.1+cu113
Is debug build: False
CUDA used to build PyTorch: 11.3
ROCM used to build PyTorch: N/A
OS: Amazon Linux 2 (x86_64)
GCC version: (GCC) 7.3.1 20180712 (Red Hat 7.3.1-15)
Clang version: Could not collect
CMake version: version 2.8.12.2
Libc version: glibc-2.26
Python version: 3.8.15 (default, Nov 16 2022, 23:39:29) [GCC 7.3.1 20180712 (Red Hat 7.3.1-15)] (64-bit runtime)
Python platform: Linux-4.14.296-222.539.amzn2.x86_64-x86_64-with-glibc2.2.5
Is CUDA available: True
CUDA runtime version: 11.7.99
GPU models and configuration: GPU 0: NVIDIA A10G
Nvidia driver version: 515.65.01
cuDNN version: Could not collect
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
Versions of relevant libraries:
[pip3] mypy-extensions==0.4.3
[pip3] numpy==1.23.3
[pip3] open-clip-torch==2.7.0
[pip3] pytorch-lightning==1.7.6
[pip3] torch==1.12.1+cu113
[pip3] torchdiffeq==0.2.3
[pip3] torchmetrics==0.11.0
[pip3] torchsde==0.2.5
[pip3] torchvision==0.13.1+cu113
[conda] No relevant packages
If I was sure this wouldn't degrade performance for people who don't have this problem, I'd be happy to open a PR — contributors feel free to offer guidance!
