update sth. at 2025-01-16 22:04:20

README.md

@@ -11,3 +11,23 @@ short_description: jina-embeddings-v3
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+
+## Usage
+
+You can generate embeddings by sending a POST request to one of the following endpoints:
+
+- `/generate_embeddings`
+- `/api/v1/embeddings`
+- `/hf/v1/embeddings`
+- `/api/v1/chat/completions`
+- `/hf/v1/chat/completions`
+
+Example request using `curl`:
+
+```sh
+curl -X POST https://sanbo1200-jina-embeddings-v3.hf.space/api/v1/embeddings \
+  -H "Content-Type: application/json" \
+  -d '{
+    "input": "Your text string goes here",
+    "model": "jinaai/jina-embeddings-v3"
+  }'

app.py

@@ -5,10 +5,14 @@ import torch
 from typing import List, Dict
 import uvicorn
 
-# 定义响应模型
+# 定义请求和响应模型
+class EmbeddingRequest(BaseModel):
+    input: str
+    model: str = "jinaai/jina-embeddings-v3"
+
 class EmbeddingResponse(BaseModel):
     status: str
-    embeddings: List[Listfloat]]
+    embeddings: List[List[float]]
 
 # 创建FastAPI应用
 app = FastAPI(
@@ -22,41 +26,33 @@ model_name = "jinaai/jina-embeddings-v3"
 tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
 model = AutoModel.from_pretrained(model_name, trust_remote_code=True)
 
-async def generate_embeddings(text: str):
+@app.post("/generate_embeddings", response_model=EmbeddingResponse)
+@app.post("/api/v1/embeddings", response_model=EmbeddingResponse)
+@app.post("/hf/v1/embeddings", response_model=EmbeddingResponse)
+@app.post("/api/v1/chat/completions", response_model=EmbeddingResponse)
+@app.post("/hf/v1/chat/completions", response_model=EmbeddingResponse)
+async def generate_embeddings(request: EmbeddingRequest):
     try:
         # 使用分词器处理输入文本
-        inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512)
-        
- #生成嵌入
+        inputs = tokenizer(request.input, return_tensors="pt", truncation=True, max_length=512)
+
+        # 生成嵌入
         with torch.no_grad():
             embeddings = model(**inputs).last_hidden_state.mean(dim=1)
-        
- return EmbeddingResponse(
+
+        return EmbeddingResponse(
             status="success",
             embeddings=embeddings.numpy().tolist()
         )
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))
 
-@app.post("/api/v1/embeddings")
-@app.post("/hf/v1/embeddings")
-async def embedding(request: Request):
-    try:
-        data = await request.json()
-        text = data.get('input', '')
-        if not text:
-            raise HTTPException(status_code=400, detail="Input text is missing")
-        
- return await generate_embeddings(text)
-    except Exception as e:
-        raise HTTPException(status_code=500, detail=str(e))
-
 @app.get("/")
 async def root():
     return {
         "status": "active",
         "model": model_name,
-        "usage": "Send POST request to /api/v1/embeddings"
+        "usage": "Send POST request to /generate_embeddings or /api/v1/embeddings or /hf/v1/embeddings"
     }
 
 if __name__ == "__main__":