search

PyImageSearch / Deep-Learning-for-Computer-Vision-with-Python

2 stars 1 forks source link

Which python version is reccomended for tensorflow 1.12. #26

Open yohannjs opened 1 year ago

yohannjs commented 1 year ago

Are you using one of our development environments?

No

If you are using your own development environment then please provide the following details.

Describe the issue

Bundle: ImageNet Chapter: 15: Training a Faster R-CNN From Scratch

I am currently trying to install the TFOD API according to the instructions in the "Installing the TensorFlow Object Detection API" blog-post. But, it does not specify which python version that should be used. This is not arbitrary as not all python versions support the version of tensorflow that is recommended for download (1.12). Python 3.11 did not support installation of TF 1.12 at least. So it would be great if you could specify which version of python you recommend, and also provide a guide of how to download alternate versions of Python as this too is not an easy task.

Any other information

yohannjs commented 1 year ago

I see now in that you specified that you were using python 3.6 in the guide, So I switched to python 3.6.10. But, I am struggling to download some of the packages listed in the guide. Specifically, I get errors when I am building the wheels for

Could you specify which versions of the following packages that you outline in the Installing the TensorFlow Object Detection API" blog-post:

FIRST

numpy
scipy
cython
scikit-learn
matplotlib
lxml
jupyter
Pillow
imutils

SECOND

tensorflow==1.12

THIRD

beautifulsoup4

FOURTH

opencv-contrib-python

SIXTH

pycocotools

yohannjs commented 1 year ago

I also get an error when trying to install the COCO-API from the github repository image

vikaskookna commented 1 year ago

Hi, @yohannjs TensorFlow 1.12 is compatible with Python 2.7 and Python 3.4-3.7. It's generally recommended to use the latest version of Python that is compatible with TensorFlow. For TensorFlow 1.12, that would be Python 3.7. However, I would recommend you use our development environments for a smooth experience.