DeepLabCut / DeepLabCut-live

SDK for running DeepLabCut on a live video stream
https://elifesciences.org/articles/61909
Other
190 stars 50 forks source link

ERROR: Could not find a version that satisfies the requirement tensorflow==1.13.1 (from deeplabcut-live) #35

Closed arge-7 closed 3 years ago

arge-7 commented 3 years ago

Hello, I'm running into some strange version issues trying to install on Nvidia Jetson Nano. I've followed the instructions to the letter or so it seems, until the very end when I ran into the issue of "ModuleNotFoundError: No module named 'check_install'". I saw the explanation of the very issue in the resolved issues page which got me to looking into what version of deeplabcut-live I had. I installed via pip install deeplabcut-live as instructed, but it seems to have installed version 0.0.0. I cannot figure out why it won't install version 0.0.3 since it's what is on PyPI. If I enter pip install deeplabcut-live -vv into terminal I get:

Using pip 21.0 from /home/rj/dlc-live/lib/python3.6/site-packages/pip (python 3.6) Non-user install because user site-packages disabled Created temporary directory: /tmp/pip-ephem-wheel-cache-q1umg_hg Created temporary directory: /tmp/pip-req-tracker-vu8r4nvi Initialized build tracking at /tmp/pip-req-tracker-vu8r4nvi Created build tracker: /tmp/pip-req-tracker-vu8r4nvi Entered build tracker: /tmp/pip-req-tracker-vu8r4nvi Created temporary directory: /tmp/pip-install-t44o1y09 Requirement already satisfied: deeplabcut-live in ./dlc-live/lib/python3.6/site-packages (0.0) Requirement already satisfied: tqdm in ./dlc-live/lib/python3.6/site-packages (from deeplabcut-live) (4.56.0) Requirement already satisfied: pillow in ./dlc-live/lib/python3.6/site-packages (from deeplabcut-live) (8.1.0) Requirement already satisfied: numpy in ./dlc-live/lib/python3.6/site-packages (from deeplabcut-live) (1.16.1) Requirement already satisfied: colorcet in ./dlc-live/lib/python3.6/site-packages (from deeplabcut-live) (2.0.6) Requirement already satisfied: py-cpuinfo==5.0.0 in ./dlc-live/lib/python3.6/site-packages (from deeplabcut-live) (5.0.0) Requirement already satisfied: pandas in ./dlc-live/lib/python3.6/site-packages (from deeplabcut-live) (1.1.5) Requirement already satisfied: tables in ./dlc-live/lib/python3.6/site-packages (from deeplabcut-live) (3.6.1) Requirement already satisfied: ruamel.yaml in ./dlc-live/lib/python3.6/site-packages (from deeplabcut-live) (0.16.12) Requirement already satisfied: pyct>=0.4.4 in ./dlc-live/lib/python3.6/site-packages (from colorcet->deeplabcut-live) (0.4.8) Requirement already satisfied: param>=1.7.0 in ./dlc-live/lib/python3.6/site-packages (from colorcet->deeplabcut-live) (1.10.1) Requirement already satisfied: pytz>=2017.2 in ./dlc-live/lib/python3.6/site-packages (from pandas->deeplabcut-live) (2020.5) Requirement already satisfied: python-dateutil>=2.7.3 in ./dlc-live/lib/python3.6/site-packages (from pandas->deeplabcut-live) (2.8.1) Requirement already satisfied: six>=1.5 in ./dlc-live/lib/python3.6/site-packages (from python-dateutil>=2.7.3->pandas->deeplabcut-live) (1.15.0) Requirement already satisfied: ruamel.yaml.clib>=0.1.2 in ./dlc-live/lib/python3.6/site-packages (from ruamel.yaml->deeplabcut-live) (0.2.2) Requirement already satisfied: numexpr>=2.6.2 in ./dlc-live/lib/python3.6/site-packages (from tables->deeplabcut-live) (2.7.2) Created temporary directory: /tmp/pip-unpack-kufb3fw7 Removed build tracker: '/tmp/pip-req-tracker-vu8r4nvi'

It's odd that it's persisting on version 0.0.0 so I tried to force install with deeplabcut-live==0.0.3 and I am getting the following errors: ERROR: Could not find a version that satisfies the requirement tensorflow==1.13.1 (from deeplabcut-live) ERROR: No matching distribution found for tensorflow==1.13.1

I do not have tensorflow 1.13.1. I check with pip show tensorflow and have version 1.15.4+nv20.12 (as per the instructions to download the NVidia version < 2.0).

Thanks in advance for any guidance you may have!

arge-7 commented 3 years ago

After lots of trial and error, I tried the pip option --no-deps so that it wouldn't look at the tensorflow wheels throwing all the errors, and it worked. DLC test ran fine. I don't know where the requirement of tensorflow==1.13.1 is coming from which is still a mystery.