Instructions to use Weikaijie/HealthPulse-Qwen2.5-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Weikaijie/HealthPulse-Qwen2.5-7B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Weikaijie/HealthPulse-Qwen2.5-7B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Weikaijie/HealthPulse-Qwen2.5-7B")
model = AutoModelForCausalLM.from_pretrained("Weikaijie/HealthPulse-Qwen2.5-7B", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Weikaijie/HealthPulse-Qwen2.5-7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Weikaijie/HealthPulse-Qwen2.5-7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Weikaijie/HealthPulse-Qwen2.5-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Weikaijie/HealthPulse-Qwen2.5-7B

SGLang

How to use Weikaijie/HealthPulse-Qwen2.5-7B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Weikaijie/HealthPulse-Qwen2.5-7B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Weikaijie/HealthPulse-Qwen2.5-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Weikaijie/HealthPulse-Qwen2.5-7B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Weikaijie/HealthPulse-Qwen2.5-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Weikaijie/HealthPulse-Qwen2.5-7B with Docker Model Runner:
```
docker model run hf.co/Weikaijie/HealthPulse-Qwen2.5-7B
```

本草医疗 Qwen2.5-7B（HealthPulse-Qwen2.5-7B）

基于 Qwen2.5-7B-Instruct 在 Huatuo（华佗）中文医患数据集上 SFT + LoRA 微调的医疗问答大模型。聚焦医学准确性与安全性，适用于症状咨询、健康科普、导诊建议、用药常识问答等场景。

模型信息

项目	说明
基座模型	Qwen2.5-7B-Instruct
微调方法	SFT + LoRA（rank=8, alpha=16, lr=5e-5）
训练框架	LLaMA-Factory
训练数据	Huatuo（华佗）医患数据集 9,000+ 条（Alpaca 格式）
训练平台	华为云 A800（80GB）GPU
上下文长度	8K（可扩展）
语言	中文

快速使用

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
    "Weikaijie/HealthPulse-Qwen2.5-7B",
    torch_dtype="auto",
    device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained("Weikaijie/HealthPulse-Qwen2.5-7B")

messages = [{"role": "user", "content": "高血压患者日常饮食需要注意什么？"}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(out[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

vLLM（推荐，OpenAI 兼容）

pip install vllm
vllm serve Weikaijie/HealthPulse-Qwen2.5-7B --port 8000

from openai import OpenAI
client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")
resp = client.chat.completions.create(
    model="HealthPulse-Qwen2.5-7B",
    messages=[{"role": "user", "content": "最近失眠多梦，怎么调理？"}],
)
print(resp.choices[0].message.content)

下载权重

# 海外直连
hf download Weikaijie/HealthPulse-Qwen2.5-7B --local-dir ./merged_model

# 国内镜像加速
export HF_ENDPOINT=https://hf-mirror.com
hf download Weikaijie/HealthPulse-Qwen2.5-7B --local-dir ./merged_model

from huggingface_hub import snapshot_download
snapshot_download("Weikaijie/HealthPulse-Qwen2.5-7B", local_dir="./merged_model")

评测（5 维 NLU）

维度	说明	相对通用模型
医学准确性	是否符合医学共识	提升约 10%
安全性	是否不乱开药/给危险建议	同步提升
专业性	是否体现医生角色	持平/略优
完整性	是否覆盖关键信息	略低于最新国产大模型
清晰度	表达是否流畅	略低于最新国产大模型

评测方式：LLM 初审 + 医学论坛评分 + 争议样本医学人士复核；如实记录短板，完整性与清晰度较最新国产大模型低 10%-20%（基座模型代差所致）。

训练详情

数据管线：华佗原始医患数据 → Hive 暂存 + PySpark 清洗 → Alpaca JSON 格式；Dify 工作流 + GPT-4o-mini 多维打分（≥6.0 分，保留 90%-95%）
数据划分：7,000 train / 1,000 val / 1,000 test
训练参数：LoRA rank=8, alpha=16, dropout=0.1, lr=5e-5, batch=8, epochs 3-5