IlaAI-v3 🌱

Earth · Crop · Intelligence

An open-source Agricultural AI for Bharat's Farmers

---

🌾 What is IlaAI?

IlaAI (इला — Sanskrit for "the earth that gives") is an open-source LLM series built specifically for Indian agriculture. Our mission is simple — give every farmer in India access to expert agricultural advice, for free, in their own language.

"For the hands that feed a billion 🌾"

---

🏆 What's New in v3?

IlaAI-v3 is our biggest leap yet — trained on real farmer data from India's Kisan Call Center (KCC).

• 🏆 Best val loss ever — 0.598
• 🌾 175,000+ real farmer Q&A pairs from KCC
• 🗣️ Telugu support (experimental)
• ⚡ Fast responses with enable_thinking=False
• 🇮🇳 India-specific — real crops, real problems, real solutions

📊 How v3 Compares

Feature	v1	v1.1	v3
Dataset	Synthetic 96K	Thinking 100K	Real KCC 175K
Val Loss	0.695	0.821	0.598
English	✅ Good	✅ Better	✅ Excellent
Hindi	❌ Weak	⚠️ Basic	✅ Good
Telugu	❌ None	❌ None	⚠️ Experimental
Real Farmer Data	❌	❌	✅
Thinking Disabled	❌	❌	✅
Response Speed	Slow	Slow	Fast

---

🌾 About the KCC Dataset

The Kisan Call Center (KCC) is a Government of India initiative launched in 2004. Farmers across India call a toll-free number (1800-180-1551) to get expert advice on agriculture. Every call is logged — the farmer's question and the expert's answer.

This dataset contains 175,000+ real farmer questions with expert answers, covering:

• Crop diseases and pest management
• Fertilizer and pesticide recommendations
• Irrigation and water management
• Government schemes and subsidies
• Market prices and crop selection
• Region-specific advice across all Indian states

Training on this data means IlaAI speaks like a real agricultural expert — because it learned from real agricultural experts. 🌱

---

🚀 Quick Start

from mlx_lm import load, generate
from mlx_lm.sample_utils import make_sampler

model, tokenizer = load("Ila-AI/IlaAI-v3")

messages = [
    {"role": "system", "content": "You are IlaAI, an expert agricultural assistant for Indian farmers. Always respond in the same language the user writes in. Be concise, practical and helpful."},
    {"role": "user", "content": "My wheat crop has yellow spots on leaves. What should I do?"}
]

text = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    tokenize=False,
    enable_thinking=False  # Always add this for fast responses!
)

sampler = make_sampler(temp=0.3, top_p=0.9)
response = generate(model, tokenizer, prompt=text, max_tokens=500, sampler=sampler, verbose=True)

---

⚡ Important — Always Disable Thinking Mode

text = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    tokenize=False,
    enable_thinking=False  # Critical! Always add this.
)

Without enable_thinking=False, the model wastes tokens on internal reasoning and responses are slower and lower quality.

---

📊 Training Details

Detail	Value
Base Model	IlaAI-v1.1 (Qwen3-4B)
Framework	MLX LoRA
Hardware	Apple M4 Mac Mini (24GB)
Dataset 1	KCC 175K real farmer Q&A
Dataset 2	KissanAI Thinking-climate-100k
Dataset 3	1,515 Telugu Q&A pairs
Training iters	3,000
LoRA rank	8
Final Val Loss	0.598
Peak Memory	8.3 GB

---

🗣️ Language Support

Language	Status
English	✅ Excellent
Hindi	✅ Good
Telugu	⚠️ Experimental
Tamil	🔄 Coming in v4
Kannada	🔄 Coming in v4
Marathi	🔄 Coming in v4
Bengali	🔄 Coming in v4
Gujarati	🔄 Coming in v4
Punjabi	🔄 Coming in v4

---

📈 Model Evolution

Version	Dataset	Val Loss	Status
v1	KissanAI 96K synthetic	0.695	✅ Released
v1.1	KissanAI 100K thinking	0.821	✅ Released
v3	KCC 175K real + Telugu	0.598	✅ Released
v4	Full 22+ Indian languages	TBD	🔄 Coming

---

🗺️ Complete Roadmap

✅ Phase 1 — Foundation (Done)

• IlaAI-v1 — English agriculture advisory (96K synthetic rows)
• IlaAI-v1.1 — Improved English (100K thinking dataset)
• IlaAI-v3 — KCC real farmer data + Telugu experimental

🔄 Phase 2 — True Multilingual (v4)

• 22+ Indian languages with proper quality
• Real multilingual agriculture Q&A
• BPCC + AI4Bharat datasets
• Release IlaAI-v4

👁️ Phase 3 — Vision (IlaAI-Vision)

• Crop disease detection from photos
• Support all major Indian crops
• Combine text + vision into one model

📱 Phase 4 — Mobile App (Free Forever)

• Free Android app
• Free iOS app
• On-device inference (no internet needed)
• Voice input in Indian languages
• Available on Play Store & App Store

🌍 Phase 5 — Global

• Expand beyond India
• Support farmers worldwide — Africa, Southeast Asia, South America
• Open API for developers
• WhatsApp bot integration

---

🤝 Contributing

We welcome contributions from developers, farmers, agronomists, and language experts!

• 📂 Add regional crop data
• 🌐 Help with Indian language translations
• 🐛 Report issues
• 💡 Suggest features

---

📜 License

Apache 2.0 — free to use, fine-tune, and build upon.

---

🙏 Acknowledgements

• KissanAI for open-sourcing Dhenu models and datasets
• Kisan Call Center — Government of India
• Qwen Team for Qwen3 base models
• Apple MLX Team for MLX framework
• Every Indian farmer who inspired this project 🌾

---

For the hands that feed a billion 🌾
HuggingFace · GitHub