Hugging Face Inference API - 免费测试数千个开源模型

pip install requests
# 或使用官方库
pip install huggingface_hub

import requests

API_URL = "https://api-inference.huggingface.co/models/meta-llama/Llama-3.1-8B-Instruct"
headers = {"Authorization": "Bearer YOUR_HF_TOKEN"}

def query(payload):
    response = requests.post(API_URL, headers=headers, json=payload)
    return response.json()

# 调用 API
output = query({
    "inputs": "请介绍一下人工智能的发展历史。",
    "parameters": {
        "max_new_tokens": 500,
        "temperature": 0.7
    }
})

print(output)

from huggingface_hub import InferenceClient

# 初始化客户端
client = InferenceClient(token="YOUR_HF_TOKEN")

# 文本生成
response = client.text_generation(
    "请介绍一下人工智能的发展历史。",
    model="meta-llama/Llama-3.1-8B-Instruct",
    max_new_tokens=500,
    temperature=0.7
)

print(response)

from huggingface_hub import InferenceClient

client = InferenceClient(token="YOUR_HF_TOKEN")

# 流式文本生成
for token in client.text_generation(
    "写一首关于春天的诗",
    model="meta-llama/Llama-3.1-8B-Instruct",
    max_new_tokens=200,
    stream=True
):
    print(token, end="", flush=True)

curl https://api-inference.huggingface.co/models/meta-llama/Llama-3.1-8B-Instruct \
  -X POST \
  -H "Authorization: Bearer YOUR_HF_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "inputs": "请介绍一下人工智能的发展历史。",
    "parameters": {
      "max_new_tokens": 500,
      "temperature": 0.7
    }
  }'

curl https://api-inference.huggingface.co/models/stabilityai/stable-diffusion-2-1 \
  -X POST \
  -H "Authorization: Bearer YOUR_HF_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "inputs": "a beautiful sunset over the ocean, oil painting style"
  }' \
  --output image.jpg

import { HfInference } from '@huggingface/inference'

const hf = new HfInference('YOUR_HF_TOKEN')

async function generateText() {
  const result = await hf.textGeneration({
    model: 'meta-llama/Llama-3.1-8B-Instruct',
    inputs: '请介绍一下人工智能的发展历史。',
    parameters: {
      max_new_tokens: 500,
      temperature: 0.7
    }
  })
  
  console.log(result.generated_text)
}

generateText()

import time
from requests.exceptions import HTTPError

def call_api_with_rate_limit(payload):
    """带速率限制处理的 API 调用"""
    max_retries = 5
    retry_delay = 1
    
    for attempt in range(max_retries):
        try:
            response = requests.post(API_URL, headers=headers, json=payload)
            response.raise_for_status()
            return response.json()
        except HTTPError as e:
            if e.response.status_code == 429:
                # 速率限制，指数退避重试
                wait_time = retry_delay * (2 ** attempt)
                print(f"速率限制，等待 {wait_time} 秒...")
                time.sleep(wait_time)
            else:
                raise
    
    raise Exception("达到最大重试次数")

def batch_inference(inputs, batch_size=10):
    """批量推理，减少请求次数"""
    results = []
    
    for i in range(0, len(inputs), batch_size):
        batch = inputs[i:i+batch_size]
        payload = {"inputs": batch}
        
        try:
            response = requests.post(API_URL, headers=headers, json=payload)
            results.extend(response.json())
            
            # 避免速率限制
            time.sleep(1)
        except Exception as e:
            print(f"批次 {i//batch_size} 失败: {e}")
    
    return results

from huggingface_hub import InferenceClient

client = InferenceClient(token="YOUR_HF_TOKEN")

models = [
    "meta-llama/Llama-3.1-8B-Instruct",
    "mistralai/Mistral-7B-Instruct-v0.2",
    "Qwen/Qwen2.5-7B-Instruct"
]

def compare_models(prompt):
    """对比多个模型的输出"""
    results = {}
    
    for model in models:
        print(f"\n测试模型: {model}")
        try:
            response = client.text_generation(
                prompt,
                model=model,
                max_new_tokens=200
            )
            results[model] = response
            print(f"回答: {response[:100]}...")
        except Exception as e:
            results[model] = f"错误: {e}"
    
    return results

# 使用示例
prompt = "请用一句话解释什么是人工智能。"
results = compare_models(prompt)

for model, response in results.items():
    print(f"\n{model}:")
    print(response)

from huggingface_hub import InferenceClient

client = InferenceClient(token="YOUR_HF_TOKEN")

def summarize_document(document_text):
    """生成文档摘要"""
    # 使用摘要模型
    summary = client.summarization(
        document_text,
        model="facebook/bart-large-cnn",
        max_length=150,
        min_length=50
    )
    
    return summary['summary_text']

# 使用示例
document = """
人工智能（Artificial Intelligence，AI）是计算机科学的一个分支，
旨在创建能够执行通常需要人类智能的任务的系统。这些任务包括视觉感知、
语音识别、决策制定和语言翻译等。AI 的发展经历了多个阶段...
"""

summary = summarize_document(document)
print(f"摘要: {summary}")

from huggingface_hub import InferenceClient
from PIL import Image

client = InferenceClient(token="YOUR_HF_TOKEN")

def classify_image(image_path):
    """图像分类"""
    with open(image_path, "rb") as f:
        data = f.read()
    
    result = client.image_classification(
        data,
        model="google/vit-base-patch16-224"
    )
    
    return result

# 使用示例
image_path = "photo.jpg"
results = classify_image(image_path)

print("图像分类结果:")
for item in results:
    print(f"- {item['label']}: {item['score']:.2%}")