allo-/virtual_webcam_background

pip install error on TensorFlow

PeterKW opened this issue · 6 comments

PeterKW commented

pip install -r requirements.txt
ERROR: Could not find a version that satisfies the requirement tensorflow (from versions: none)
ERROR: No matching distribution found for tensorflow

Anything I should try to fix this?

allo- commented

You probably use a python versions that is (no longer) supported by tensorflow.

You may try to find which tensorflow / tensorflow-gpu / tensorflow-cpu / tensorflow-nightly packages may be available. I suppose its something like you're still using python2 and trying to install tensorflow, which only provides python3 packages (like almost every package by now).

PeterKW commented

Thank you, in the end it had to specify to use python3 ->

python3.9 -m pip install tensorflow

python3.9 -m pip install -r requirements.txt

PeterKW commented

Any ideas on how to fix the dependency conflicts?

I tried : python3 -m pip install -I numpy==1.19.3 but it didn't seem to help?

python3 -m pip install -r requirements.txt

Collecting tensorflow
  Using cached tensorflow-2.5.0-cp39-cp39-win_amd64.whl (422.6 MB)
INFO: pip is looking at multiple versions of tfjs-graph-converter to determine which version is compatible with other requirements. This could take a while.
INFO: pip is looking at multiple versions of <Python from Requires-Python> to determine which version is compatible with other requirements. This could take a while.
INFO: pip is looking at multiple versions of numpy to determine which version is compatible with other requirements. This could take a while.
ERROR: Cannot install -r requirements.txt (line 1), numpy==1.18.5 and tensorflow==2.6.0 because these package versions have conflicting dependencies.

The conflict is caused by:
    The user requested numpy==1.18.5
    tensorflow 2.6.0 depends on numpy~=1.19.2
    The user requested numpy==1.18.5
    tensorflow 2.5.1 depends on numpy~=1.19.2
    The user requested numpy==1.18.5
    tensorflow 2.5.0 depends on numpy~=1.19.2

To fix this you could try to:
1. loosen the range of package versions you've specified
2. remove package versions to allow pip attempt to solve the dependency conflict

ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/user_guide/#fixing-conflicting-dependencies
allo- commented

Try to just install the latest versions. The versions in requirements.txt are tested, but using at least a more recent numpy should be no problem (not so sure for tensorflow, which seems to change more often).

atisDu commented

Just comment out the numpy line in requirements.txt and tensorflow will automatically pull in the right numpy.

allo- commented

I wonder if we should have numpy only as indirect dependency. The project uses it directly, but I guess both tensorflow and mediapipe will always pull a compatible version in.