You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Gemma-3n-E4B KCC LoRA (Pilot)

Model Description

This repository contains LoRA adapters fine-tuned on the Kisan Call Centre (KCC) farmer advisory dataset from the Government of India.
This is a pilot / test training run intended to validate the data pipeline, training stability, and domain suitability.

⚠️ This is NOT a final production model.

Developed by

Uralstech (research prototype)

Funded by

Not externally funded (independent research)

Training Details

Base model: google/gemma-3n-E4B
Method: LoRA fine-tuning (language + attention + MLP layers)
Dataset: Kisan Call Centre (KCC) advisory transcripts
Samples: ~121,756
Epochs: 1 (pilot)
Final training loss: ~2.7

Intended Use

Research, benchmarking, and evaluation of agricultural advisory assistants.

Limitations

Pilot run only
Not optimized for mobile inference
No multilingual or safety alignment tuning

License

Gemma License. Only LoRA adapters are distributed.

Downloads last month: -; Downloads are not tracked for this model. How to track

Model tree for uralstech/gemma-3n-e4b-kcc-lora-pilot

Base model

google/gemma-3n-E4B

Adapter

(1)

this model