Ghost2513/qwen-medical-reasoning

This is a fine-tuned Qwen model for medical reasoning based on the FreedomIntelligence/medical-o1-reasoning-SFT dataset.

Model Details

Model Description

This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.

• Developed by: Ghost2513
• Model type: Causal LM
• Language(s) (NLP): English
• License: MIT License
• Finetuned from model: Qwen base model

Uses

Direct Use

Use for medical reasoning, generating step-by-step reasoning for clinical questions.

Out-of-Scope Use

Not for real-life medical decision making. Not to replace professional advice.

Bias, Risks, and Limitations

This model may contain biases from the training dataset. Users should not rely on it for clinical decisions without validation.

How to Get Started

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

tokenizer = AutoTokenizer.from_pretrained("{full_repo_name}")
model = AutoModelForCausalLM.from_pretrained("{full_repo_name}").to("cuda")

question = "How can I identify and manage early signs of sepsis in adults?"
prompt = f"{{question}}\\nReasoning: <insert reasoning here>\\nAnswer:"
inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, max_new_tokens=500)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
Training Details
Training Data: FreedomIntelligence/medical-o1-reasoning-SFT

Training Regime: Fine-tuned on Qwen base model

Hardware: GPU-based

Compute Region: [More Information Needed]

Citation
java
Copy code
@misc{qwen-medical-reasoning,
  title={Qwen Fine-tuned Medical Reasoning Model},
  author={Ghost2513},
  year={2025},
  howpublished={https://huggingface.co/{full_repo_name}}
}
"""

with open("README.md", "w") as f:
f.write(readme_text)

api.upload_file(
path_or_fileobj="README.md",
path_in_repo="README.md",
repo_id=full_repo_name,
token=HF_TOKEN
)
print("Model and README successfully uploaded to Hugging Face Hub!")

vbnet
Copy code

This script does the following:

1. Logs in to Hugging Face using `HF_TOKEN`.
2. Loads your fine-tuned Qwen model and tokenizer.
3. Creates the repo if it doesn’t exist.
4. Pushes both the model and tokenizer.
5. Generates a **full README/model card** with all sections you provided (most fields are placeholders for you to fill).
6. Uploads the README to the repository.  

---