README.md

---
library_name: transformers
license: apache-2.0
base_model: Qwen/Qwen3-4B
pipeline_tag: text-generation
language:
  - bn
  - en
tags:
  - math
  - bengali
  - reasoning
  - sft
datasets:
  - dipta007/Ganit
---

# GanitLLM-4B_SFT

[![Paper](https://img.shields.io/badge/arXiv-Paper-red)](https://arxiv.org/)
[![Dataset](https://img.shields.io/badge/HuggingFace-Dataset-yellow)](https://huggingface.co/datasets/dipta007/Ganit)
[![Models](https://img.shields.io/badge/HuggingFace-Models-orange)](https://huggingface.co/collections/dipta007/ganitllm)

## Highlights

**GanitLLM-4B_SFT** is a Bengali mathematical reasoning model trained with Supervised Fine-Tuning on the GANIT dataset. This model serves as the foundation for further RL training (GRPO/CGRPO). Key improvements over the base Qwen3-4B model:

- **+4.80 accuracy** on Bn-MGSM benchmark (69.20 → 74.00)
- **+4.10 accuracy** on Bn-MSVAMP benchmark (70.50 → 74.60)
- **86.65% Bengali reasoning** (vs 14.79% for base model)
- **80.5% fewer words** in generated solutions (943 → 184 words)

> **Note**: This is the SFT-only checkpoint. For best results, use the RL-enhanced versions: [GanitLLM-4B_SFT_CGRPO](https://huggingface.co/dipta007/GanitLLM-4B_SFT_CGRPO) or [GanitLLM-4B_SFT_GRPO](https://huggingface.co/dipta007/GanitLLM-4B_SFT_GRPO).

## Model Overview

| Property | Value |
|----------|-------|
| **Model Type** | Causal Language Model |
| **Base Model** | Qwen/Qwen3-4B |
| **Parameters** | 4B |
| **Training** | Supervised Fine-Tuning |
| **Context Length** | 4,096 tokens |
| **Language** | Bengali, English |

## Training Details

This model was trained with a single-stage pipeline:

1. **Supervised Fine-Tuning (SFT)**: Trained on GANIT-SFT (~11k examples) to ground reasoning in Bengali

### Training Data
- **Dataset**: GANIT-SFT (11,023 examples)
- **Format**: Bengali math problems with chain-of-thought reasoning
- **Structure**: `<think>` tags for reasoning, `<answer>` tags for final answer

## Quickstart

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "dipta007/GanitLLM-4B_SFT"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
    device_map="auto"
)

problem = "একটি দোকানে ১২টি আপেল আছে। যদি ৫টি আপেল বিক্রি হয়, তাহলে কতটি আপেল বাকি থাকবে?"

prompt = f"""A conversation takes place between the user and the assistant. The user asks a question, and the assistant solves the problem. Please reason step by step in Bengali, and put your final answer in the <answer> </answer> tags.

Question: {problem}"""

messages = [{"role": "user", "content": prompt}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

generated_ids = model.generate(**model_inputs, max_new_tokens=2048, temperature=0.7)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist()
response = tokenizer.decode(output_ids, skip_special_tokens=True)
print(response)
```

### Using vLLM

```bash
vllm serve dipta007/GanitLLM-4B_SFT --max-model-len 4096
```

## Performance

| Model | Bn-MGSM | Bn-MSVAMP | Avg. Words | Bengali % |
|-------|---------|-----------|------------|-----------|
| Qwen3-4B (base) | 69.20 | 70.50 | 943 | 14.79% |
| **GanitLLM-4B_SFT** | **74.00** | **74.60** | **184** | **86.65%** |

## Related Models

| Model | Parameters | Training | Link |
|-------|------------|----------|------|
| GanitLLM-4B_SFT_CGRPO | 4B | SFT + CGRPO | [Link](https://huggingface.co/dipta007/GanitLLM-4B_SFT_CGRPO) |
| GanitLLM-4B_SFT_GRPO | 4B | SFT + GRPO | [Link](https://huggingface.co/dipta007/GanitLLM-4B_SFT_GRPO) |
| **GanitLLM-4B_SFT** | 4B | SFT | [Link](https://huggingface.co/dipta007/GanitLLM-4B_SFT) |
| GanitLLM-4B_CGRPO | 4B | CGRPO | [Link](https://huggingface.co/dipta007/GanitLLM-4B_CGRPO) |

## Citation

```bibtex
will be updated
```

## License

This model is released under the Apache 2.0 License.