---
license: mit
datasets:
- openlifescienceai/medmcqa
language:
- en
base_model:
- Qwen/Qwen2.5-0.5B-Instruct
tags:
- grpo,
- biomed,reasoning
---

# Model Card for BioXP

BioXP-0.5B is a 🤗 Transformers-based model trained using our two-stage fine-tuning approach:

1. Supervised Fine-Tuning (SFT): The model was initially fine-tuned on labeled data(MedMCQA) to achieve strong baseline accuracy on multiple-choice medical QA tasks.

2. Group Relative Policy Optimization (GRPO): In the second stage, GRPO was applied to further align the model with human-like reasoning patterns.
This reinforcement learning technique enhances the model’s ability to generate coherent, high-quality explanations and improve answer reliability.

## Model Details

### Model Description

This model is a finetuned version of Qwen/Qwen2.5-0.5B-Instruct, a 0.5 billion parameter language model from the Qwen2 family. 
The finetuning was performed using reinforcement learning approach: Group Relative Policy Optimization (GRPO).

- **Developed by:** Qwen (original model), finetuning by rgb2gbr
- **Funded by :** rgb2bgr
- **Shared by :** rgb2bgr
- **Model type:** Causal Language Model
- **Language(s) (NLP):** English
- **License:** MIT
- **Finetuned from model:** Qwen/Qwen2.5-0.5B-Instruct


## Uses

The model fine-tuned using grpo method on openlifescienceai/medmcqa, can answer and identify the correct options in dataset with 44.02% accuracy


### Out-of-Scope Use

This model should not be used for generating harmful, biased, or inappropriate content. It's important to be aware of the potential limitations and biases inherited from the base model and the finetuning data.

## Bias, Risks, and Limitations

As a large language model, this model may exhibit biases present in the training data. The finetuning process may have amplified or mitigated certain biases. Further evaluation is needed to understand the full extent of these biases and limitations.

## How to Get Started with the Model

You can load this model using the `transformers` library in Python:

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "rgb2gbr/BioXP-0.5B-MedMCQA" 
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", torch_dtype="auto")

prompt = "Identify the right answer and elaborate it"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=200)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))