Increasing Training Resources - Konnsy/REAML2022-hackathon GitHub Wiki

Depending on how your approach is designed, your own GPU or CPU will be sufficient for training. If you want to alleviate this resource constraint, you can use services like https://colab.research.google.com to increase your training speed. With this service, computing power of idling servers can be used for free (with a Google account). Remember to extract the code and the corresponding trained model and store them locally on your machine.

Code frame of task 1 in Google Colab
You can access the code in https://colab.research.google.com/drive/11MVhsk07QDCtFH5yhQLxUEV1qS1q9ELc?usp=sharing At first, add the file to your account in order to make it persistable.

An alternative way can be found at colab_example.ipynb

When using Colab
Make sure that GPU usage is set. To activate, select Runtime -> Change runtime type -> select GPU as Hardware accelerator otherwise only CPU will be used which will slow down the training process.