Google Colab integration

[rvlb]

No description provided.

[codecov]

Codecov Report

Merging #22 (df3b8a9) into main (163c06c) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main      #22   +/-   ##
=======================================
  Coverage   61.22%   61.22%           
=======================================
  Files          13       13           
  Lines        1212     1212           
=======================================
  Hits          742      742           
  Misses        470      470           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 163c06c...df3b8a9. Read the comment docs.

[fjsj]

• Could you please make a Colab for notebooks/End-to-End-Matching-Example.ipynb?
• Please also check all links inside the Colab notebooks.

[fjsj]

Also, in the examples section in readme, please make a table like this with colab links: https://github.com/KevinMusgrave/pytorch-metric-learning/blob/master/examples/README.md

[fjsj]

One more thing: is there a way to force Colab to use a GPU? I had to manually use "Change runtime type" to use a GPU.

[rvlb]

Not that I've found, pytorch-metric-learning also asks that we do it manually

Our readme: https://github.com/vintasoftware/entity-embed/blob/5f3b34e8b42bbed4f621a1c5a37edf7de75e5509/notebooks/google-colab/README.md

[fjsj]

• Please keep the current descriptions we have in README for the examples.
• Make a Colab for notebooks/End-to-End-Matching-Example.ipynb. Not sure how to load the model though, maybe we'll have to pre-generate it. Please research if possible to have disk persistence in Colab across Notebooks.
• If it's not possible to programatically request a GPU, then please add the instruction to the top of the Colab Notebook.

[rvlb]

Persistence seems to be possible via Google Drive integration https://stackoverflow.com/a/66670818

[fjsj]

Persistence seems to be possible via Google Drive integration

In this case, we can leave the code that saves to Drive commented and put an explanation on the Notebook.

[rvlb]

Please note that these urls consider the files already exist on main.

[fjsj]

LGTM