Skip to content

This project is an implementation of fine-tuning the Gemma 2b-it model on a custom dataset and deploy the model on Android.

License

Notifications You must be signed in to change notification settings

NSTiwari/Gemma-on-Android

Repository files navigation

Gemma on Android

This project is an implementation of fine-tuning the Gemma 2b-it model on a custom dataset and deploy the fine-tuned model on Android

Pipeline:

Demo Output:

Steps to run:

  1. Clone the repository on your local machine.
  2. Open the project in Android Studio.
  3. Edit InferenceModel.kt file on Line 44 by replacing YOUR_MODE_NAME.bin with the actual name of your model.
  4. Build the project.
  5. Install the Android app on your phone and enjoy using SciGemma.

Resources:

  1. Follow along three blog series explaining the code in detail:

    Part 1: Step-by-Step Dataset Creation- Unstructured to Structured

    Part 2: Fine Tune - Gemma 2b-it model

    Part 3: Deploying SciGemma on Android

  2. Fine-tuned model on 🤗: https://huggingface.co/NSTiwari/fine_tuned_science_gemma2b-it

  3. Try our model on HFSpaces: https://huggingface.co/spaces/Aashi/NSTiwari-fine_tuned_science_gemma2b-it?logs=container

  4. Checkout demo video on YouTube: https://www.youtube.com/watch?v=T_HDsVHTrwg

Acknowledgment:

This project was developed during Google's ML Developer Programs Gemma sprint. We thank the MLDP team for the opportunity.

About

This project is an implementation of fine-tuning the Gemma 2b-it model on a custom dataset and deploy the model on Android.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published