I'm studying image segmentation using tensorflow.
This time I tried using the oxford dataset.
macos Google Colab
I finished writing the code once with vscode, but I still have the problem that the processing is slow because I can not use the GPU.
Therefore, I decided to use Google Colab published by Google teacher. You can use the GPU for free here.
First, I pasted my code, but it says that the Oxford dataset has not been downloaded. That's strange. I was able to do it normally with VScode, but when I looked it up, I found out. ["Image segmentation by u-net"](https://data-analysis-stats.jp/%E6%B7%B1%E5%B1%9E%E5%AD%A6%E7%BF%92/u- net% E3% 81% AB% E3% 82% 88% E3% 82% 8B% E7% 94% BB% E5% 83% 8F% E3% 82% BB% E3% 82% B0% E3% 83% A1% E3% 83% B3% E3% 83% 86% E3% 83% BC% E3% 82% B7% E3% 83% A7% E3% 83% B3image-segmentation% E3% 81% AE% E8% A7% A3% E8% AA% AC /) I found this article, so I used it as a reference.
!pip install -q git+https://github.com/tensorflow/examples.git
!pip install -q -U tfds-nightly
It seemed like I had to write these code first.
Apparently, even if you downloaded it once on the terminal of the pc, when you use it with google colab, it seems that you need to download it again on colab.
And after all, GPU was able to process faster than CPU!
Recommended Posts