| --- |
| library_name: transformers |
| tags: [fine-tuning, medical, oncology, reasoning, chain-of-thought, Qwen-1.5B] |
| --- |
| |
| # Model Card for Medical Oncology Reasoning Fine-Tuned Model |
|
|
| This is a fine-tuned version of the **DeepSeek-R1-Distill-Qwen-1.5B** model, specifically adapted for medical oncology reasoning tasks using chain-of-thought prompting. The fine-tuning was performed on a curated subset of the [FreedomIntelligence/medical-o1-reasoning-SFT](https://huggingface.co/datasets/FreedomIntelligence/medical-o1-reasoning-SFT) dataset. This model is designed to provide detailed, step-by-step reasoning when answering medical questions, making it suitable for clinical decision support, medical education, and research. |
|
|
| ## Model Details |
|
|
| ### Model Description |
|
|
| This model leverages the capabilities of **DeepSeek-R1-Distill-Qwen-1.5B** and has been fine-tuned to enhance its performance in medical reasoning. It incorporates chain-of-thought prompting to produce detailed explanations for clinical queries. The fine-tuning process focused on tasks related to clinical diagnostics, treatment planning, and medical reasoning. |
|
|
| - **Developed by:** Arihant Tripathi |
| - **Model type:** Causal Language Model (LLM) |
| - **Language(s):** English |
| - **Finetuned from model:** [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B) |
|
|
| ### Model Sources |
|
|
| - **Repository:** [EndOfLe/OncoFineTuned](https://huggingface.co/EndOfLe/OncoFineTuned) |
|
|
| ## Uses |
|
|
| ### Direct Use |
|
|
| This model can be used directly to generate detailed, step-by-step medical reasoning in response to clinical queries. It is especially useful for: |
| - Medical diagnosis support. |
| - Clinical reasoning and treatment planning. |
| - Medical education and research. |
|
|
| ### Downstream Use |
|
|
| The model can be integrated into larger clinical decision support systems or educational tools that require natural language understanding and detailed reasoning for medical queries. |
|
|
| ### Out-of-Scope Use |
|
|
| The model is not intended for: |
| - Replacing expert medical advice or making final clinical decisions. |
| - General-purpose language generation without domain adaptation. |
| - High-stakes applications where errors could have severe consequences without expert oversight. |
|
|
| ## Bias, Risks, and Limitations |
|
|
| The model’s output should be treated as an aid to human decision-making rather than a substitute for professional medical advice. Key considerations include: |
| - The model may generate outdated or incorrect medical information. |
| - The reasoning is based on the training data and might reflect its inherent biases. |
| - Use in clinical settings should always involve human review and validation. |
|
|
| ### Recommendations |
|
|
| Users should verify the model’s outputs with expert medical knowledge and ensure its use complies with clinical standards and ethical guidelines. |
|
|
| ## How to Get Started with the Model |
|
|
| You can get started with the model using the following code snippet: |
|
|
| ```python |
| from transformers import AutoTokenizer, AutoModelForCausalLM |
| |
| model_name = "EndOfLe/OncoFineTuned" |
| tokenizer = AutoTokenizer.from_pretrained(model_name) |
| model = AutoModelForCausalLM.from_pretrained(model_name) |
| |
| input_text = "What are the causes of colon cancer?" |
| inputs = tokenizer(input_text, return_tensors="pt") |
| outputs = model.generate(**inputs, max_new_tokens=256) |
| print(tokenizer.decode(outputs[0], skip_special_tokens=True)) |
| |
