Recommend Colab + GPUs for transfer learning episode

[qualiaMachine]

The pretrained model(s) we use in the transfer learning episode are quite large (relative to prev. examples in workshop at least). To speed things up, I'm going to recommend my learners use Google Colab for that episode, with the GPU enabled. I added this recommendation as a "spoiler" block so it doesn't take up too much of the episode real estate, for those that want to stick to the local setup. Those who wish to use Colab should follow the instructions added to the spoiler block.

On a related note, I will be adding another exercise to this lesson to explore finetuning. This is another motivating reason for why I really think a GPU is called for here. Plus, this is an opportunity to explain the importance of GPUs to learners.

[github-actions]

Thank you!

Thank you for your pull request 😃

🤖 This automated message can help you check the rendered files in your submission for clarity. If you have any questions, please feel free to open an issue in {sandpaper}.

If you have files that automatically render output (e.g. R Markdown), then you should check for the following:

• 🎯 correct output
• 🖼️ correct figures
• ❓ new warnings
• ‼️ new errors

Rendered Changes

🔍 Inspect the changes: https://github.com/carpentries-lab/deep-learning-intro/compare/md-outputs..md-outputs-PR-583

The following changes were observed in the rendered markdown documents:

 5-transfer-learning.md | 62 +++++++++++++++++++++++++++++++++++++++++++++++++-
 md5sum.txt             |  2 +-
 2 files changed, 62 insertions(+), 2 deletions(-)

What does this mean?

If you have source files that require output and figures to be generated (e.g. R Markdown), then it is important to make sure the generated figures and output are reproducible.

This output provides a way for you to inspect the output in a diff-friendly manner so that it's easy to see the changes that occur due to new software versions or randomisation.

⏱️ Updated at 2025-05-15 20:04:00 +0000

[carschno]

I am very hesitant to actively recommend proprietary products as part of the teaching, even though I understand the advantages. We also refer to Binder instead, although we need to make sure it serves the purpose.
Anyway, running things locally would remain my preference, and is probably becoming a smaller problem.