dhananjayyy23
/

shikshaedge-gemma4-e4b

Model card Files Files and versions

ShikshaEdge — Gemma 3 4B fine-tuned for Indian K-12

Fine-tuned from Gemma 3 4B using QLoRA (Unsloth) on 2,170 curriculum-aligned Q&A pairs from NCERT textbooks in Hindi, Marathi, and English. Class 6-10 curriculum.

Use case

Offline AI tutoring for Indian K-12 students via Ollama.

Training

Method: QLoRA (r=16) via Unsloth
Dataset: 2,170 NCERT Q&A pairs
Languages: Hindi, Marathi, English
Hardware: Google Colab T4
Epochs: 3

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

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for dhananjayyy23/shikshaedge-gemma4-e4b

Base model

google/gemma-3-4b-pt

Finetuned

google/gemma-3-4b-it

Finetuned

(733)

this model