---
license: apache-2.0
language:
- en
- mr
base_model:
- NousResearch/Hermes-3-Llama-3.2-3B
library_name: adapter-transformers
pipeline_tag: text-generation
tags:
- trl
- sft
---
# Koyna-V1-3b-instruct (Under Development)
## Fruit Farming Assistance Model (Bilingual: Marathi and English)


A comprehensive AI model designed to assist fruit farmers, particularly those working with grapes, by providing accurate answers to common farming-related questions in both **Marathi** and **English**. This model is fine-tuned on a synthetic dataset of question-answer pairs, covering topics such as crop management, pest control, irrigation, and harvesting techniques.

## Model Overview

This model uses advanced natural language processing (NLP) techniques to provide relevant answers to agricultural questions in **Marathi** and **English**. It can help farmers by offering advice on best practices, responding to common queries about grape farming, and providing information on various farming techniques.

- **Model type**: Fine-tuned Language Model
- **Framework**: PyTorch, Transformers
- **Fine-tuned on**: Synthetic dataset of question-answer pairs related to fruit farming, with a focus on grapes
- **Languages Supported**: Marathi, English
- **Input format**: Text (Questions from farmers)
- **Output format**: Text (Answers to the questions)

## Installation

To use this model, you will need to install the `transformers` library and PyTorch.

```bash
pip install transformers
pip install torch
```

## Usage

Here’s how you can load and use the model to answer farming-related questions in both Marathi and English:

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Load the pre-trained model and tokenizer from Hugging Face
model_name = "Govind222/Koyna-V1-3b-instruct"
model = AutoModelForCausalLM.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)

# Example question in Marathi
question = "द्राक्षांसाठी कॅल्शियम नायट्रेट कधी वापरावं?"

# Tokenize the question
inputs = tokenizer(question, return_tensors="pt")

# Get the model's answer
with torch.no_grad():
    answer = model.generate(inputs['input_ids'])

# Decode the answer from the tokens
answer_text = tokenizer.decode(answer[0], skip_special_tokens=True)
print("Answer:", answer_text)
```

## Example Questions for Fruit Farming (Marathi and English)

You can ask the model farming-related questions in Marathi or English, such as:

### Example Questions in Marathi

1. **द्राक्षांसाठी कॅल्शियम नायट्रेट कधी वापरावं?**
2. **द्राक्षांसाठी बुरशीजन्य रोग नियंत्रण कधी करावं?**
3. **द्राक्षांसाठी सर्वोत्तम तापमान कोणते आहे?**
4. **द्राक्ष पिकासाठी पाणी किती प्रमाणात द्यावे?**
5. **द्राक्ष वेलीला जास्त थंडी असल्यास, काय उपाय करावा?**
6. **द्राक्ष वेलीला फळांचा आकार छोटा होत आहे, यावर काय उपाय करावा?**

## Training Details

This model was fine-tuned using a synthetic dataset containing question-answer pairs. These pairs cover a wide range of topics related to grape farming and other fruit farming practices, and the dataset is in both Marathi and English. The dataset was created to simulate real-world queries that farmers may ask, focusing on improving practical knowledge for daily agricultural tasks.

## Model Details

- **Model ID**: Govind222/Koyna-V1-3b-instruct
- **Version**: 1.0.0
- **Architecture**: [e.g., BERT, RoBERTa, etc.]
- **License**: Apache-2.0

## Citation

If you use or reference this model, please cite it as follows:

```bibtex
@misc{Koyna,
  author = {Govind Barbade},
  title = {Koyna-V1-3b-instruct},
  year = {2024},
  url = {https://huggingface.co/Govind222/Koyna-V1-3b-instruct}
}
```

## License

This model is licensed under the Apache 2.0 License.

## Acknowledgments

The synthetic dataset was created using a combination of publicly available farming resources input in both Marathi and English.
Special thanks to the farming community for helping to refine and validate the model's capabilities.