Model Card for dhee-chat-mistral-mr

A fine-tuned Marathi conversational model based on mistralai/Mistral-7B-v0.3 , optimized for Marathi language understanding and generation.

Model Details

• Base Model: Mistral 7B v0.3
• Fine-tuning Method: LoRA (Low-Rank Adaptation)
• Dataset: ai4bharat/indic-align
• Language: Marathi
• Model ID: dheeyantra/dhee-chat-mistral-mr

Intended Uses & Limitations

This model is intended for use in Marathi conversational applications, such as chatbots and virtual assistants. As it is fine-tuned on the ai4bharat/indic-align dataset, its knowledge and conversational style are primarily shaped by this data.

Limitations:

• The model's responses are based on the patterns and information present in the training data. It may generate incorrect or biased information.
• Performance may vary depending on the complexity and nuance of the input.
• The model is primarily focused on Marathi and may not perform well in other languages or code-mixed scenarios unless explicitly trained for them.

How to Get Started with Hugging Face Transformers

You can use the following Python code to load and run inference with the dheeyantra/dhee-chat-mistral-mr model:

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_path = "dheeyantra/dhee-chat-mistral-mr"

# Load tokenizer and model
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(model_path)

# Move model to GPU if available
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

# Prepare chat messages
messages = [
    {"role": "User", "content": "किती वेद आहेत?"},
    {"role": "Dhee", "content": "ऋग्वेद, यजुर्वेद, सामवेद आणि अथर्ववेद असे चार वेद आहेत."},
    {"role": "User", "content": "ऋग्वेदा बद्दल अधिक सांगा?"}
]

# Apply chat template to get prompt
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)

# Tokenize prompt
inputs = tokenizer(prompt, return_tensors="pt").to(device)

# Generate output
with torch.no_grad():
    output_ids = model.generate(
        **inputs,
        max_new_tokens=64,
        do_sample=True,
        temperature=0.7,
        top_p=0.95,
        pad_token_id=tokenizer.eos_token_id
    )

# Decode generated text
generated_text = tokenizer.decode(output_ids[0][inputs['input_ids'].shape[-1]:], skip_special_tokens=True)
results = [{"generated_text": generated_text}]

print("Generated text:")
print(results[0]['generated_text'])

Disclaimer

This model is provided as-is. Users should be aware of its potential limitations and biases before deploying it in any application. Responsible AI practices should be followed.

Training Configuration

The model was fine-tuned using the following LoRA and training parameters:

LoRA Parameters:

• r: 16
• target_modules: ["q_proj", "k_proj", "v_proj", "o_proj", "gate_proj", "up_proj", "down_proj"]
• lora_alpha: 16
• lora_dropout: 0
• bias: "none"
• use_gradient_checkpointing: "unsloth"
• use_rslora: False
• loftq_config: None

Training Arguments:

• gradient_accumulation_steps: 4
• warmup_ratio: 0.03
• fp16: True
• optim: "adamw_8bit"
• max_seq_length: 32768

Acknowledgements

We extend our sincere gratitude to the following organizations for their invaluable contributions to this project:

• NxtGen: For generously providing the necessary infrastructure that powered the model training.
• AI4Bharat: For developing and making available the indic-align dataset, which was crucial for fine-tuning this model.

Citation

If you use this model in your research or applications, please cite:

@misc{dheenxtgen2025,
  title={ dhee-chat-mistral-mr : A Compact Language Model for Marathi},
  author={Dheeyantra Research Labs},
  year={2025},}
}