---
base_model: unsloth/Qwen3-8B-unsloth-bnb-4bit
library_name: peft
pipeline_tag: text-generation
tags:
- base_model:adapter:unsloth/Qwen3-8B-unsloth-bnb-4bit
- lora
- sft
- transformers
- trl
- unsloth
datasets:
- krisfu/delicate_medical_r1_data_chinese
language:
- zh
---

# Model Card for Model ID

<!-- Provide a quick summary of what the model is/does. -->
This is a LoRA adapter for `unsloth/Qwen3-8B-unsloth-bnb-4bit`, fine-tuned for Chinese medical question-answering tasks. It's designed to provide helpful, detailed, and harmless responses to general medical inquiries based on the provided dataset.



## Model Details

### Model Description

<!-- Provide a longer summary of what this model is. -->
This model is the result of Supervised Fine-Tuning (SFT) on the Qwen3-8B base model using the PEFT library with LoRA. The fine-tuning was performed using the Unsloth library to optimize for speed and memory efficiency. The model is specifically trained on the `krisfu/delicate_medical_r1_data_chinese` dataset, which contains a collection of Chinese medical questions and answers.

- **Developed by:** Ronan
- **Model type:** Qwen3 (Transformer-based Language Model)
- **Language(s) (NLP):** Chinese (zh)
- **License:** Apache-2.0 (Inherited from the base model, but you can choose another)
- **Finetuned from model:** `unsloth/Qwen3-8B-unsloth-bnb-4bit`


### Model Sources [optional]

<!-- Provide the basic links for the model. -->
- **Repository:** [https://huggingface.co/ronan7878/yiliao_qwen/]
- **Base Model Repository:** [https://huggingface.co/unsloth/Qwen3-8B-unsloth-bnb-4bit](https://huggingface.co/unsloth/Qwen3-8B-unsloth-bnb-4bit)


## Uses

<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->

### Direct Use

<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
This model is intended for direct use as a question-answering assistant in the Chinese medical domain. It can be used in chatbots, as a research aid, or for generating preliminary medical information. **This model is not a substitute for professional medical advice, diagnosis, or treatment.**

**Example prompt:**
"我最近总是感到疲劳，并且伴有头痛，可能是什么原因？"

### Out-of-Scope Use

This model should **not** be used for:
* Providing definitive medical diagnoses or treatment plans.
* Emergency medical situations.
* Generating content that could be harmful, unethical, or misleading.
* Any use case where model error could lead to harm.

## Bias, Risks, and Limitations

The model's knowledge is limited to the data it was trained on, which may not be comprehensive or up-to-date. It may generate plausible-sounding but incorrect medical information. The training data may contain inherent biases, which the model could reproduce or amplify. It is a language model and has no true understanding of medical concepts.

### Recommendations

Always consult a qualified healthcare professional for any medical concerns. The model's outputs should be fact-checked and reviewed by a human expert before being used in any critical application.

## How to Get Started with the Model

Use the code below to get started with the model using the Unsloth library.

```python
from unsloth import FastLanguageModel
import torch

model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = "ronan7878/yiliao_qwen",
    max_seq_length = 2048,
    dtype = None,
    load_in_4bit = True,
)

# Perform inference
inputs = tokenizer(
[
    "我最近总是感到疲劳，并且伴有头痛，可能是什么原因？"
], return_tensors = "pt").to("cuda")

outputs = model.generate(**inputs, max_new_tokens = 128, use_cache = True)
print(tokenizer.batch_decode(outputs))
```

## Training Details

### Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

The model was fine-tuned on the krisfu/delicate_medical_r1_data_chinese dataset. This dataset consists of Chinese medical question-and-answer pairs designed for fine-tuning language models.

### Training Procedure

<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
The model was trained using the Supervised Fine-Tuning (SFT) method with the TRL library. The Unsloth library was used to enable 4-bit quantization and LoRA for efficient training.


#### Training Hyperparameters

- Training regime: bnb-4bit quantization with LoRA adapters

- LoRA r: 16

- LoRA alpha: 32

- Optimizer: AdamW

- Learning Rate: 2e-4

### Framework versions

- PEFT 0.17.1
- Transformers 4.43.3
- TRL 0.9.6
- Unsloth 2025.9