outdated requirements?

[xenogenesi]

I am trying to run it locally but I have problems with dependency versions, I have tried with Debian/sid and then with docker (which is based on nvidia/cuda:10.1-cudnn7-devel-ubuntu18.04 and is the same image for the face-alignment project which works)

ERROR: Could not find a version that satisfies the requirement opencv-python==4.1.0.25 (from -r requirements.txt (line 4)) (from versions: 3.4.8.29, 3.4.9.31, 3.4.9.33, 3.4.10.35, 3.4.10.37, 3.4.11.39, 3.4.11.41, 3.4.11.43, 3.4.11.45, 4.1.2.30, 4.2.0.32, 4.2.0.34, 4.3.0.36, 4.3.0.38, 4.4.0.40, 4.4.0.42, 4.4.0.44, 4.4.0.46)
ERROR: No matching distribution found for opencv-python==4.1.0.25 (from -r requirements.txt (line 4))

ERROR: Could not find a version that satisfies the requirement tensorflow==1.12.0 (from -r requirements.txt (line 5)) (from versions: 2.2.0rc1, 2.2.0rc2, 2.2.0rc3, 2.2.0rc4, 2.2.0, 2.2.1, 2.3.0rc0, 2.3.0rc1, 2.3.0rc2, 2.3.0, 2.3.1, 2.4.0rc0)
ERROR: No matching distribution found for tensorflow==1.12.0 (from -r requirements.txt (line 5))

ERROR: Could not find a version that satisfies the requirement torch==1.1.0 (from -r requirements.txt (line 6)) (from versions: 0.1.2, 0.1.2.post1, 0.1.2.post2, 1.4.0, 1.5.0, 1.5.1, 1.6.0, 1.7.0)
ERROR: No matching distribution found for torch==1.1.0 (from -r requirements.txt (line 6))

ERROR: Could not find a version that satisfies the requirement torchvision==0.3.0 (from -r requirements.txt (line 7)) (from versions: 0.1.6, 0.1.7, 0.1.8, 0.1.9, 0.2.0, 0.2.1, 0.2.2, 0.2.2.post2, 0.2.2.post3, 0.5.0, 0.6.0, 0.6.1, 0.7.0, 0.8.0, 0.8.1)
ERROR: No matching distribution found for torchvision==0.3.0 (from -r requirements.txt (line 7))

I tried to upgrade the packages with the nearest available stable version

--- a/requirements.txt
+++ b/requirements.txt
@@ -1,9 +1,9 @@
 librosa==0.7.0
 numpy==1.17.1
 opencv-contrib-python==4.2.0.34
-opencv-python==4.1.0.25
-tensorflow==1.12.0
-torch==1.1.0
-torchvision==0.3.0
+opencv-python==4.1.2.30
+tensorflow==2.2.0
+torch==1.4.0
+torchvision==0.6.0
 tqdm==4.45.0
 numba==0.48

But then numpy fail to build

    ...
    [cut]
    ...
    numpy/random/entropy.c:23684:15: error: ‘__Pyx_PyCode_New’ undeclared (first use in this function); did you mean ‘PyCode_New’?
         py_code = __Pyx_PyCode_New(
                   ^~~~~~~~~~~~~~~~
                   PyCode_New
    error: Command "gcc -pthread -B /opt/conda/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE=1 -D_LARGEFILE64_SOURCE=1 -DNPY_NO_DEPRECATED_API=0 -Inumpy/core/include -Ibuild/src.linux-x86_64-3.8/numpy/core/include/numpy -Inumpy/core/src/common -Inumpy/core/src -Inumpy/core -Inumpy/core/src/npymath -Inumpy/core/src/multiarray -Inumpy/core/src/umath -Inumpy/core/src/npysort -I/opt/conda/include/python3.8 -Ibuild/src.linux-x86_64-3.8/numpy/core/src/common -Ibuild/src.linux-x86_64-3.8/numpy/core/src/npymath -Ibuild/src.linux-x86_64-3.8/numpy/core/src/common -Ibuild/src.linux-x86_64-3.8/numpy/core/src/npymath -c numpy/random/entropy.c -o build/temp.linux-x86_64-3.8/numpy/random/entropy.o -MMD -MF build/temp.linux-x86_64-3.8/numpy/random/entropy.o.d -U__GNUC_GNU_INLINE__ -std=c99 -msse2" failed with exit status 1
    ----------------------------------------
  Rolling back uninstall of numpy
  Moving to /opt/conda/bin/f2py
   from /tmp/pip-uninstall-tyzmyngj/f2py
  Moving to /opt/conda/bin/f2py3
   from /tmp/pip-uninstall-tyzmyngj/f2py3
  Moving to /opt/conda/bin/f2py3.8
   from /tmp/pip-uninstall-tyzmyngj/f2py3.8
  Moving to /opt/conda/lib/python3.8/site-packages/numpy-1.19.2.dist-info/
   from /opt/conda/lib/python3.8/site-packages/~umpy-1.19.2.dist-info
  Moving to /opt/conda/lib/python3.8/site-packages/numpy/
   from /opt/conda/lib/python3.8/site-packages/~umpy
ERROR: Command errored out with exit status 1: /opt/conda/bin/python -u -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'/tmp/pip-install-mkks_ft8/numpy/setup.py'"'"'; __file__='"'"'/tmp/pip-install-mkks_ft8/numpy/setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' install --record /tmp/pip-record-c67btjrp/install-record.txt --single-version-externally-managed --compile --install-headers /opt/conda/include/python3.8/numpy Check the logs for full command output.

Any hint to get it working?

[prajwalkr]

Hello, I cannot help you regarding the NumPy build, but can you try with the right Python version (3.6 or 3.7) and let me know if you fail to install the requirements for that as well?

[xenogenesi]

Thanks, I saw that the numpy compilation seems to be a python 3.8 specific problem, as soon as I have time I will try to install the 3.6 in the docker image (there is a good chance that the 3.7 will work or should I stay on 3.6 to be sure?)

[prajwalkr]

Most people have tried with 3.6. I think you have a better chance with 3.6.

[xenogenesi]

Ok, it seems to work with the dependencies specified in requirements, if you are interested here is a gist of the Dockerfile (to clean and fix for cuda which does not work for me for the version in the card too old) and .dockerignore.

Thanks.

[prajwalkr]

If you can add it with instructions on how to use and create a pull request, I can merge it into the main repo. I have no experience with docker, so I would not be able to contribute here.

Glad your issue is solved! :)

[xenogenesi]

Thanks, sorry if I answer here even if the issue is closed, I cleaned up the gist a bit and added some instructions in the header, it is easy to use once a compatible version of docker is installed, the commands must be given in the project directory , the first build of the docker image takes some time (and some disk space) to download / install / build but only needs to be done once, docker run ... is used to instantiate the image and access it via the shell. I just tried inference.py and it works! Other modules, such as for training, may not find one or more of the dependencies and fail, the Dockerfile should be updated in that case, installing the modules and rebuilding the image.

I think right now it is still incomplete to open a PR, but it could be a good starting point if someone with docker experience and a gpu suitable for training wants to try.

[PraffulVarshney]

@prajwalkr @snehitvaddi @dipam7 @xenogenesi
Hello, I am Prafful Varshney, a Sophomore at IIT Roorkee working on a course ML project Lip-Sync. I have some doubts regarding this model. Can we have a small meeting on google meet or any other platform? Please give us time when you are free.
Looking forward to receiving an optimistic response.
Any help would be highly appreciated.
Thank You
Regards

[SohamGovande]

I hope this message finds you well. I encountered a similar issue as the original poster when using Python 3.10, so I switched to Python 3.6 as recommended. However, I have run into another problem when attempting to install the dependencies using pip install -r requirements.txt.

The installation process seems to get stuck at the following step for an extended period (over an hour):

Building wheel for opencv-contrib-python (PEP 517) ...

I have tried searching for any existing solutions within the repository's issues and related discussions, but I haven't found any relevant information so far. I was wondering if anyone else has experienced this issue and if there's a known solution or workaround for it.

Thanks!

[kkkiqjn]

global cap.cpp:597 cv::VideoWriter::open VIDEOIO(CV_IMAGES): raised OpenCV exception:

OpenCV(4.7.0) D:\a\opencv-python\opencv-python\opencv\modules\videoio\src\cap_images.cpp:253: error: (-5:Bad argument) CAP_IMAGES: can't find starting number (in the name of file): temp/result.avi in function 'cv::icvExtractPattern'

100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [04:28<00:00, 268.58s/it]
Traceback (most recent call last):
File "C:\Users\saket\Wav2Lip-GFPGAN\Wav2Lip-master\inference.py", line 280, in
main()
File "C:\Users\saket\Wav2Lip-GFPGAN\Wav2Lip-master\inference.py", line 277, in main
subprocess.call(command, shell=platform.system() != 'Windows')
File "C:\Users\saket\anaconda3\lib\subprocess.py", line 349, in call
with Popen(*popenargs, **kwargs) as p:
File "C:\Users\saket\anaconda3\lib\subprocess.py", line 951, in init
self._execute_child(args, executable, preexec_fn, close_fds,
File "C:\Users\saket\anaconda3\lib\subprocess.py", line 1420, in _execute_child
hp, ht, pid, tid = _winapi.CreateProcess(executable, args,
FileNotFoundError: [WinError 2] The system cannot find the file specified

I am getting this , can any one solve this

[dizhenx]

ne solve this

Have you solve it?

[bryanlimktscv]

Hope this might help someone:

I had installed Python 3.6.8 on my Windows machine and was running pip install --timeout=360 -r requirements.txt.

Firstly, make sure to use the 64-bit Python version as PyTorch does NOT work on the 32-bit Python version (Reference: PyTorch Windows Installation).

I was facing this issue:

ERROR: Could not find a version that satisfies the requirement torch==1.1.0 (from versions: none)
ERROR: No matching distribution found for torch==1.1.0

This was resolved by going to: PyTorch CPU Wheel Files
(Note: For CUDA 10.0, use PyTorch CUDA 10.0 Wheel Files.
For CUDA 9.0, use PyTorch CUDA 9.0 Wheel Files).

Then use Ctrl+F to search for:

torch-1.1.0-cp{CPython version}-cp{CPython version}m-win_amd64.whl

Replace {CPython version} with your Python version, for example 36 for Python 3.6.

Then download that file. Install using:

pip install <path to wheel file>

(Reference: Stack Overflow)

---