Closed SoftologyPro closed 6 months ago
noted.. will push a fix shortly !
noted.. will push a fix shortly !
OK, thanks. Can you let me know when it is ready to test?
@SoftologyPro can you check the green-head branch and let me know if it’s working for you? (Also delete the comfy_runner folder inside Dough,if it’s present, it will create a new one automatically)
@SoftologyPro can you check the green-head branch and let me know if it’s working for you? (Also delete the comfy_runner folder inside Dough,if it’s present, it will create a new one automatically)
How do I git clone the "green-head branch"?
Hi @SoftologyPro,
You need to fetch and switch to the branch by running this in your Terminal:
git fetch git checkout green-head
Hi @SoftologyPro,
You need to fetch and switch to the branch by running this in your Terminal:
git fetch git checkout green-head
D:\Tests\Dough>git fetch D:\Tests\Dough>git checkout green-head error: pathspec 'green-head' did not match any file(s) known to git
Do one thing for now, go into the data_repo.py file and find this method - acquire_lock.. and add the line "res = None" as shown in the image.. then start the app
OK, that gets past the error. There is still this warning at startup,
Please replace st.experimental_get_query_params with st.query_params.
st.experimental_get_query_params will be removed after 2024-04-11.
Refer to our [docs page](https://docs.streamlit.io/library/api-reference/utilities/st.query_params) for more information.
but it seems to start OK after that. I start the James Dean watercolor example, the models downloaded and various packages are installed.
Then after a while this error...
DEBUG:app_logger:database busy, retrying
error occured: No module named 'websocket'
Traceback (most recent call last):
File "D:\Tests\Dough\banodoco_runner.py", line 298, in check_and_update_db
output = predict_gpu_output(data['workflow_input'], data['file_path_list'], \
File "D:\Tests\Dough\utils\ml_processor\gpu\utils.py", line 23, in predict_gpu_output
from comfy_runner.inf import ComfyRunner
File "D:\Tests\Dough\comfy_runner\inf.py", line 10, in <module>
import websocket
ModuleNotFoundError: No module named 'websocket'
DEBUG:app_logger:database busy, retrying
DEBUG:app_logger:database busy, retrying
Ah.. I think you didn’t install websocket… close the app, run “pip install websocket” and then start the app.. although weirdly you shouldn’t be seeing “database busy” message… I feel like you missed some steps or didn’t install the packages in venv
Ah.. I think you didn’t install websocket… close the app, run “pip install websocket” and then start the app.. although weirdly you shouldn’t be seeing “database busy” message… I feel like you missed some steps or didn’t install the packages in venv
OK, typo in your requirements. You have "websockets" and not "websocket".
pip install websocket
fixes that.
But then I get
ModuleNotFoundError: No module named 'fuzzywuzzy'
pip install fuzzywuzzy
fixes that
There is also so the message
D:\Tests\Dough\voc_dough\lib\site-packages\fuzzywuzzy\fuzz.py:11: UserWarning: Using slow pure-python SequenceMatcher. Install python-Levenshtein to remove this warning
warnings.warn('Using slow pure-python SequenceMatcher. Install python-Levenshtein to remove this warning')
so I also ran
pip install python-Levenshtein
UI starts. I click the film strip icon button "Animate 'Shot 1'" That takes a few seconds to show the 3 images. I then click the bottom "Generate Animation Clip" button.
Another error
ModuleNotFoundError: No module named 'numexpr'
fixed with
pip install numexpr
Now it stops with...
To see the GUI go to: http://127.0.0.1:4333
DEBUG:app_logger:Process 49488 (Port 4333)
DEBUG:app_logger:comfy server is running
FETCH DATA from: D:\Tests\Dough\ComfyUI\custom_nodes\ComfyUI-Manager\extension-node-map.json
FETCH DATA from: D:\Tests\Dough\ComfyUI\custom_nodes\ComfyUI-Manager\custom-node-list.json
Update check done.
DEBUG:app_logger:Updating ip-adapter-plus_sd15.bin to ip-adapter-plus_sd15.bin
DEBUG:app_logger:Updating pytorch_model.bin to SD1.5\pytorch_model.bin
DEBUG:app_logger:Updating vae-ft-mse-840000-ema-pruned.safetensors to vae-ft-mse-840000-ema-pruned.safetensors
DEBUG:app_logger:Updating anything-v3-fp16-pruned.safetensors to anything-v3-fp16-pruned.safetensors
DEBUG:app_logger:Updating v3_sd15_sparsectrl_rgb.ckpt to SD1.5\animatediff\v3_sd15_sparsectrl_rgb.ckpt
DEBUG:app_logger:Updating v3_sd15_adapter.ckpt to SD1.5\animatediff\v3_sd15_adapter.ckpt
DEBUG:app_logger:Updating v3_sd15_mm.ckpt to v3_sd15_mm.ckpt
DEBUG:app_logger:Updating v3_sd15_sparsectrl_rgb.ckpt to SD1.5\animatediff\v3_sd15_sparsectrl_rgb.ckpt
INFO:app_logger:Generating output please wait
INFO:app_logger:Error generating output WebSocket.__init__() missing 3 required positional arguments: 'environ', 'socket', and 'rfile'
DEBUG:app_logger:Process 49488 (Port 4333)
error occured: 'file_paths'
Traceback (most recent call last):
File "D:\Tests\Dough\banodoco_runner.py", line 298, in check_and_update_db
output = predict_gpu_output(data['workflow_input'], data['file_path_list'], \
File "D:\Tests\Dough\utils\ml_processor\gpu\utils.py", line 35, in predict_gpu_output
return output['file_paths'] # ignoring text output for now {"file_paths": [], "text_content": []}
KeyError: 'file_paths'
I think the very first time you mistakenly ran the app without activating the environment… I simple fix for this could be to delete the old venv folder (as you installed some packages manually), create a new venv, activate it and do “pip install -r requirements.txt” and “pip install -r comfy_runner/requirements.txt”
OK, I now see I missed the pip install websocket
on your Windows install documentation. That was my problem.
I created a quick install.bat batch file to make sure I am installing Dough cleanly from scratch.
@echo off
echo *** Deleting Dough directory if it exists
if exist Dough\. rd /S /Q Dough
echo *** Cloning Dough
git clone --depth 1 -b staging https://github.com/banodoco/Dough.git
cd Dough
echo *** setting up virtual environment
python -m venv venv
call venv\Scripts\activate.bat
echo *** updating pip
python.exe -m pip install --upgrade pip
echo *** installing requirements
pip install -r requirements.txt
pip install websocket
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
call venv\Scripts\deactivate.bat
copy .env.sample .env
cd ..
echo *** finished Dough install
pause
Then I start it with a quick run.bat
@echo off
cd Dough
call venv\scripts\activate.bat
call entrypoint.bat
call venv\scripts\deactivate.bat
cd..
And now everything works as expected.
One minor issue still. When the movie is built the stats show
Moviepy - Building video C:\Users\Jason\AppData\Local\Temp\tmpau9v7y4h.mp4.
Moviepy - Writing video C:\Users\Jason\AppData\Local\Temp\tmpau9v7y4h.mp4
Moviepy - Done !
Moviepy - video ready C:\Users\Jason\AppData\Local\Temp\tmpau9v7y4h.mp4
That movie does not exist. The movie is under the Dough sub directory
videos\17727bfd-14f9-4adc-8a29-c4f5c7e8d27d\assets\videos\1_final\
This script is pretty good! I will add something like this in the repo very soon. Regarding your issue, I didn’t get what went wrong, is the video not showing up in the app? (The temp files that are created while generation are automatically deleted).
Regarding your issue, I didn’t get what went wrong, is the video not showing up in the app? (The temp files that are created while generation are automatically deleted).
No, it is not showing in the UI. Is it supposed to display down the bottom after the "Generate Animation Clip" button? This is in Firefox.
It is supposed to be shown at the top inside the “animate shot” section.. can you check if it’s there?
It is supposed to be shown at the top inside the “animate shot” section.. can you check if it’s there?
Yes, it is up top :) I expeccted to be down the bottom and never scrolled up. All working fine now. Thanks for all your help.
Actually, not 100% right. The first run after install it gets to the end with
Desired Speed Change: 2.29
Moviepy - Building video C:\Users\Jason\AppData\Local\Temp\tmpp32wh2t4.mp4.
Moviepy - Writing video C:\Users\Jason\AppData\Local\Temp\tmpp32wh2t4.mp4
t: 0% 0/45 [00:00<?, ?it/s, now=None]
Moviepy - Done !
Moviepy - video ready C:\Users\Jason\AppData\Local\Temp\tmpp32wh2t4.mp4
no audio to sync
But when I scroll up there is no output movie displayed. Then I realised there is the little toggle to the right "show video" that is off by default. Maybe that needs to be enabled by default.
Implemented that idea in the branch that's coming tomorrow - thanks for the suggestion @SoftologyPro!
Cloned the repo. Installed all deps into a new clean virtual environment (voc_dough). When the UI appears it shows this wanring
and this error
If I create a new project, it shows that the project is created, but then the above error is displayed again.