search

idealo / imagededup

😎 Finding duplicate images made easy!
https://idealo.github.io/imagededup/
Apache License 2.0
5.15k stars 455 forks source link

Fix/installation #144

Closed tanujjain closed 2 years ago

tanujjain commented 3 years ago

Running python setup.py install on the dev branch fails.

Tensorflow and numpy aren't playing well with each other. Leave tf>1.0 in setup.py with no mention of numpy (i.e., relying on tf to get numpy) leads to an error since tf 2.4.1 gets installed along with numpy 1.20.1 (due to pip resolver algo), but tf 2.4.1 needs numpy=~1.9.2. Explicitly mentioning numpy <1.20.0, makes the installation work. Additionally, newer versions of numpy (1.20.1), scipy (1.6 onwards) and matplotlib do not support python 3.6 anymore. The changes proposed in this PR will also work from Python 3.7 onwards.

datitran commented 3 years ago

Hey looking good. Overall I believe that setting tf>1.0 is not the right approach. TensorFlow has many upward/downward compatibility problems. Could imagine that tf >= 3.0 will break everything but then we can change it accordingly to 1.0 < tf =<2.0.

tanujjain commented 3 years ago

@datitran We would no longer be able to support py3.6 after this fix is shipped. Just want to confirm if you're onboard with that?

datitran commented 3 years ago

I think that should be okay until we find a long-term fix. We should find a better solution on this.

collinarnett commented 3 years ago

@tanujjain Is this going to be merged with main any time soon?