New Updated Version Doesn't Work!

[davidraymondcross]

Stop pretending like this works. I followed all the directions and tried 20 times. This does not work. It creates the same error every time.

*** Error executing callback ui_tabs_callback for C:\Program Files\stable-diffusion\stable-diffusion-webui\extensions\SadTalker\scripts\extension.py
Traceback (most recent call last):
File "C:\Program Files\stable-diffusion\stable-diffusion-webui\modules\script_callbacks.py", line 283, in ui_tabs_callback
res += c.callback() or []
File "C:\Program Files\stable-diffusion\stable-diffusion-webui\extensions\SadTalker\scripts\extension.py", line 172, in on_ui_tabs
from app_sadtalker import sadtalker_demo
File "C:\Program Files\stable-diffusion\stable-diffusion-webui/extensions/SadTalker\app_sadtalker.py", line 3, in
from src.gradio_demo import SadTalker
File "C:\Program Files\stable-diffusion\stable-diffusion-webui/extensions/SadTalker\src\gradio_demo.py", line 6, in
from src.generate_batch import get_data
File "C:\Program Files\stable-diffusion\stable-diffusion-webui/extensions/SadTalker\src\generate_batch.py", line 8, in
import src.utils.audio as audio
File "C:\Program Files\stable-diffusion\stable-diffusion-webui/extensions/SadTalker\src\utils\audio.py", line 1, in
import librosa
File "C:\Program Files\stable-diffusion\stable-diffusion-webui\venv\lib\site-packages\librosa_init_.py", line 211, in
from . import core
File "C:\Program Files\stable-diffusion\stable-diffusion-webui\venv\lib\site-packages\librosa\core_init_.py", line 9, in
from .constantq import * # pylint: disable=wildcard-import
File "C:\Program Files\stable-diffusion\stable-diffusion-webui\venv\lib\site-packages\librosa\core\constantq.py", line 1058, in
dtype=np.complex,
File "C:\Program Files\stable-diffusion\stable-diffusion-webui\venv\lib\site-packages\numpy_init_.py", line 338, in getattr
raise AttributeError(former_attrs[attr])
AttributeError: module 'numpy' has no attribute 'complex'.
np.complex was a deprecated alias for the builtin complex. To avoid this error in existing code, use complex by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use np.complex128 here.
The aliases was originally deprecated in NumPy 1.20; for more details and guidance see the original release note at:
https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations

---

Running on local URL: http://127.0.0.1:7860

To create a public link, set share=True in launch().
Startup time: 32.3s (prepare environment: 6.4s, import torch: 5.7s, import gradio: 2.5s, setup paths: 3.9s, initialize shared: 2.9s, other imports: 1.9s, load scripts: 1.1s, create ui: 7.2s, gradio launch: 0.8s).
Creating model from config: C:\Program Files\stable-diffusion\stable-diffusion-webui\configs\v1-inference.yaml
Applying attention optimization: Doggettx... done.
Model loaded in 15.7s (load weights from disk: 11.4s, create model: 0.4s, apply weights to model: 1.9s, calculate empty prompt: 1.9s).

I followed the steps - https://github.com/OpenTalker/SadTalker/blob/main/docs/webui_extension.md

I set my webui_user.bat to "set SADTALKER_CHECKPOINTS=(my file extension I put here)"

I installed FFmpeg prior to doing this. I changed the environment variables on the PC PATH to FFmpeg

I followed all the instructions that [vinthony] wrote and all the suggestions he made to other people with my problem.

It is time VINTHONY to admit that your software is flawed and take it off the GIT Hub. It doesn't work. It may have used to work around a year ago, but it no longer works. Fix your instructions or remove your failed software.

[eoffermann]

This is inappropriate (and more than a little weird).

Lots of github projects fall out of sync with recent package updates when the original developers move on to other things. THis is especially true for academic research, and EVEN MORE SO with AI research the last few years. The people that create a research project like this usually have other research to do.

We are extremely fortunate that as part of publishing their research that people advancing technology in the space also release their models and code. We're doubly fortunate when they release it with a generous opensource license. SadTalker is free software in every sense of the word. The source is Apache 2.0 licensed, and nobody has to pay a dime for it. You can work with the software, modify the software, even advance the technology contained therein and claim patents on derivative projects.

The errors you're getting are because the libraries you're running are too recent and/or out of sync with each other. Sometimes that happens. You shouldn't come in and demand that somebody fix it for you or take the git repo offline, though. That's just silly.

If I were to guess based on your traceback, I'd say you're running the wrong version of librosa. I think if you uninstall numpy and then pip install librosa==0.9.2 it will correctly install the corresponding numpy and librosa versions you need. That said, it's fairly likely that you'll have other package versions to sort out but maybe that's the end of it.

Your post here is a good example of why each time someone else adapts the source for their environment, those dependency resolutions don't get pushed to the main project: because their specific needs may not be yours and those changes will just break things for someone else.

[oisilener1982]

Dont expect any update and fix from the devs. The main author of this project is very busy and collaborating with japanese ai researcher for another project. Sadtalker is dead but the fix is to backread a little

[davidraymondcross]

I already moved on to another working git in ComfyUI. People need to be warned. I spent over 8 hours of my life trying to install this. Nobody else should have to suffer my same fate. So anyone reading this just know there is no solution to make this work.

[oisilener1982]

You are wrong. Sadtalker is working fine in the latest version of auto 1111

[samuthub]

I'm done. (Windows11) but cannot use webui.bat

=My path=
D:\SadTalker
D:\SadTalker\checkpoints
D:\SadTalker\gfpgan\weights\

=Environment=
D:\miniconda3\envs\sadtalker\

=How to install FFMPEG=
https://www.youtube.com/watch?v=YQuFqsNi2G4

2024_06_26_00.11.50.mp4

[mustofakamal1]

Try this for each issue you have encountered:

1. np.complex error, run this in venv pip install librosa==0.9.2
2. np.float error at extensions/SadTalker/src/face3d/util/my_awing_arch.py line 18, change the code preds = preds.astype(np.float, copy=False) to preds = preds.astype(float, copy=False)
3. ValueError: setting an array element with a sequence. error at extensions/SadTalker/src/face3d/util/preprocess.py, change the code trans_params = np.array([w0, h0, s, t[0], t[1]]) to trans_params = np.array([w0, h0, s, t[0], t[1]], dtype="object")

In my case, I encountered those issues, now it works. Hope it helps.