Downloads · 30 days
4
1% of all-time downloads
shikshaedge-gemma3-gguf
dhananjayyy23/shikshaedge-gemma3-gguf
shikshaedge-gemma3-gguf is a machine learning model from dhananjayyy23. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
basemodel: google/gemma-3-4b-it language: - hi - mr - en license: apache-2.0 tags: - education - india - multilingual - gguf - ollama - gemma-3 - shikshaedge ---
All-time downloads
517
Public
Repo size
4.1 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.gguf4.1 GB · 100%
At a glance
- Access
- Public
- Created
- Apr 5, 2026
- Updated
- Apr 5, 2026
- SHA
- ecdfae3f
Model card
From the Hugging Face model README
base_model: google/gemma-3-4b-it language:
- hi
- mr
- en license: apache-2.0 tags:
- education
- india
- multilingual
- gguf
- ollama
- gemma-3
- shikshaedge
ShikshaEdge — Gemma 3 4B GGUF
Fine-tuned from Gemma 3 4B on 2,170 NCERT curriculum Q&A pairs in Hindi, Marathi, and English. Optimized for offline AI tutoring for Indian K-12 students (Class 6-10).
Run with Ollama
ollama pull hf.co/dhananjayyy23/shikshaedge-gemma3-gguf
ollama run hf.co/dhananjayyy23/shikshaedge-gemma3-gguf
Use case
- Offline AI tutoring for Class 6-10 Indian students
- Runs on Raspberry Pi 4 / low-cost hardware
- No internet required after setup
- Hindi, Marathi, English multilingual support
Training details
- Base model: Gemma 3 4B IT
- Fine-tuning: QLoRA (r=16) via Unsloth on Google Colab T4
- Dataset: 2,170 NCERT Science & Math Q&A pairs
- Languages: Hindi (7/20), English (7/20), Marathi (6/20) per batch
- Epochs: 3
- Final loss: ~2.3
Quantization
- Format: GGUF Q8_0
- Size: 4.1GB
- Compatible with Ollama, llama.cpp
Part of ShikshaEdge project
Built for the Gemma 4 Good Hackathon 2026. Dataset: ShikshaEdge NCERT Curriculum Dataset
- Created
- Apr 5, 2026
- Updated
- Apr 5, 2026