How to use from
MLX LM
Generate or start a chat session
# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "Ila-AI/IlaAI-v1.1"
Run an OpenAI-compatible server
# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "Ila-AI/IlaAI-v1.1"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8080/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "Ila-AI/IlaAI-v1.1",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'
Quick Links
IlaAI Logo

IlaAI-v1.1 ๐ŸŒฑ

Earth ยท Crop ยท Intelligence

An open-source Agricultural AI for Bharat's Farmers

HuggingFace GitHub License Base Model Val Loss


๐ŸŒพ What's New in v1.1?

IlaAI-v1.1 is an improved version of IlaAI-v1 with:

  • 5x more training data โ€” 9,000 rows vs 1,800 rows
  • Better dataset โ€” KissanAI Thinking-climate-100k (chain-of-thought reasoning)
  • More training โ€” 3,000 iterations vs 2,000 iterations
  • Better Val Loss โ€” 0.821 vs 0.874
  • Multilingual system prompt โ€” responds in user's language

๐Ÿš€ Quick Start

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

model, tokenizer = load("Ila-AI/IlaAI-v1.1")

messages = [
    {"role": "system", "content": "You are IlaAI, an expert agricultural assistant for Indian farmers. Always respond in the same language the user writes in. Keep answers concise, practical and under 200 words."},
    {"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
)

sampler = make_sampler(temp=0.7, top_p=0.9)
response = generate(model, tokenizer, prompt=text, max_tokens=1000, sampler=sampler, verbose=True)

๐Ÿ“Š Training Details

Detail Value
Base Model IlaAI-v1 (Qwen3-4B)
Framework MLX LoRA
Hardware Apple M4 Mac Mini (24GB)
Dataset KissanAI/Thinking-climate-100k
Training rows 9,000
Validation rows 1,000
Training iters 3,000
LoRA rank 8
Final Val Loss 0.821
Peak Memory 4.507 GB

๐Ÿ—ฃ๏ธ Language Support

Language Status
English โœ… Excellent
Hindi โœ… Good
Telugu โš ๏ธ Basic
Tamil โš ๏ธ Basic
Kannada โš ๏ธ Basic
Others โš ๏ธ Basic

Full multilingual support coming in IlaAI-v2 with real multilingual agriculture data.


โš ๏ธ Limitations

  • Best performance in English โ€” primary training language
  • Hindi โ€” good quality responses
  • Other Indian languages โ€” basic support, improving in v2
  • Text only โ€” vision model coming separately

๐Ÿ—บ๏ธ Roadmap

โœ… Phase 1 โ€” Foundation

  • IlaAI-v1 โ€” English agriculture advisory
  • IlaAI-v1.1 โ€” Improved English + basic multilingual

๐Ÿ”„ Phase 2 โ€” True Multilingual (v2)

  • Real multilingual agriculture data (BPCC + AI4Bharat)
  • 22+ Indian languages with proper quality
  • Release IlaAI-v2

๐Ÿ‘๏ธ Phase 3 โ€” Vision

  • Crop disease detection from photos
  • IlaAI-Vision model

๐Ÿ“ฑ Phase 4 โ€” Mobile App

  • Free Android & iOS app
  • On-device inference
  • Voice input in Indian languages

๐ŸŒ Phase 5 โ€” Global

  • Expand beyond India
  • Support farmers worldwide
  • API for developers

๐Ÿ“œ License

Apache 2.0 โ€” free to use, fine-tune, and build upon.


๐Ÿ™ Acknowledgements

  • KissanAI for open-sourcing Dhenu models and datasets
  • Qwen Team for Qwen3 base models
  • Apple MLX Team for MLX framework
  • Every Indian farmer who inspired this project ๐ŸŒพ

For the hands that feed a billion ๐ŸŒพ
HuggingFace ยท GitHub
Downloads last month
10
Safetensors
Model size
4B params
Tensor type
U32
ยท
BF16
ยท
MLX
Hardware compatibility
Log In to add your hardware

4-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for Ila-AI/IlaAI-v1.1

Finetuned
Qwen/Qwen3-4B
Adapter
(1)
this model
Adapters
1 model

Dataset used to train Ila-AI/IlaAI-v1.1