Problem description

I'm trying to run the GPT-2 model on the Termux app.

To avoid all dependency related issues, I'm mainly using a virtualized Ubuntu-like environment. Everything gets installed as expected. The model gets downloaded from its source (around 6 GB) via the library, but when the code arrives at using the model, the app crashes within a few seconds.

I believe this is because of the intensive operations the program might be trying to do.

Is there any way to get past this limitation on Android 14 assuming I have access to the device's shell via adb?

Steps to reproduce the behavior.

1. Install Ubuntu
2. Install pip in it.
3. Install transformers library via pip.
4. Write some basic code that uses transformers library to load and use the GPT2 model and just print the output.
5. Run the script.

What is the expected behavior?

The script should run, even if it takes a while to do so.

System information

• Termux application version: v0.118.0
• Android OS version: 14
• Device model: Pixel 7A