AgriAssist is a fine-tuned vision-language model (VLM) built on Qwen2-VL-7B-Instruct, designed specifically for Indian agricultural applications. It is trained on curated datasets covering major crops, weeds, pests, and diseases, enabling robust recognition and basic reasoning over agricultural images.

Features

Domain-specific fine-tuning for Indian agriculture
Recognition of crops, weeds, and pests
Instruction-tuned for multimodal reasoning
Trained on multiple public datasets: MH-Weed16, PlantVillage, AgroBench
Ready for integration in applications requiring agricultural image understanding

Usage

python

from transformers import AutoProcessor, AutoModelForCausalLM

processor = AutoProcessor.from_pretrained("your-username/AgriAssist")
model = AutoModelForCausalLM.from_pretrained("your-username/AgriAssist")

# Example usage
inputs = processor(images=image_list, text="Identify the pest in the image", return_tensors="pt")
outputs = model.generate(**inputs)

RetrO21/AgriModelLo

AgriAssist: Domain-Specific Vision-Language Model for Indian Agriculture

Features

Usage

RetrO21/
AgriModelLo