|
import os |
|
import requests |
|
from fastapi import FastAPI, HTTPException |
|
from fastapi.middleware.cors import CORSMiddleware |
|
from llama_cpp import Llama |
|
from pydantic import BaseModel |
|
import uvicorn |
|
|
|
|
|
MODEL_URL = "https://huggingface.co/unsloth/DeepSeek-R1-Distill-Qwen-1.5B-GGUF/resolve/main/DeepSeek-R1-Distill-Qwen-1.5B-Q5_K_M.gguf" |
|
MODEL_NAME = "DeepSeek-R1-Distill-Qwen-1.5B-Q5_K_M.gguf" |
|
MODEL_DIR = "model" |
|
MODEL_PATH = os.path.join(MODEL_DIR, MODEL_NAME) |
|
|
|
|
|
os.makedirs(MODEL_DIR, exist_ok=True) |
|
|
|
|
|
if not os.path.exists(MODEL_PATH): |
|
print(f"Downloading model from {MODEL_URL}...") |
|
response = requests.get(MODEL_URL, stream=True) |
|
if response.status_code == 200: |
|
with open(MODEL_PATH, "wb") as f: |
|
for chunk in response.iter_content(chunk_size=8192): |
|
f.write(chunk) |
|
print("Model downloaded successfully!") |
|
else: |
|
raise RuntimeError(f"Failed to download model: HTTP {response.status_code}") |
|
else: |
|
print("Model already exists. Skipping download.") |
|
|
|
|
|
app = FastAPI(title="DeepSeek-R1 OpenAI-Compatible API") |
|
|
|
|
|
app.add_middleware( |
|
CORSMiddleware, |
|
allow_origins=["*"], |
|
allow_methods=["*"], |
|
allow_headers=["*"], |
|
) |
|
|
|
|
|
print("Loading model...") |
|
try: |
|
llm = Llama( |
|
model_path=MODEL_PATH, |
|
n_ctx=2048, |
|
n_threads=4, |
|
n_gpu_layers=0, |
|
verbose=False |
|
) |
|
print("Model loaded successfully!") |
|
except Exception as e: |
|
raise RuntimeError(f"Failed to load model: {str(e)}") |
|
|
|
|
|
class ChatCompletionRequest(BaseModel): |
|
model: str = "DeepSeek-R1-Distill-Qwen-1.5B" |
|
messages: list[dict] |
|
max_tokens: int = 128 |
|
temperature: float = 0.7 |
|
top_p: float = 0.9 |
|
stream: bool = False |
|
|
|
|
|
class ChatCompletionResponse(BaseModel): |
|
id: str = "chatcmpl-12345" |
|
object: str = "chat.completion" |
|
created: int = 1693161600 |
|
model: str = "DeepSeek-R1-Distill-Qwen-1.5B" |
|
choices: list[dict] |
|
usage: dict |
|
|
|
|
|
@app.post("/v1/chat/completions") |
|
async def chat_completion(request: ChatCompletionRequest): |
|
try: |
|
prompt = "\n".join([f"{msg['role']}: {msg['content']}" for msg in request.messages]) |
|
prompt += "\nassistant:" |
|
|
|
|
|
response = llm( |
|
prompt=prompt, |
|
max_tokens=request.max_tokens, |
|
temperature=request.temperature, |
|
top_p=request.top_p, |
|
stop=["</s>"] |
|
) |
|
|
|
|
|
return ChatCompletionResponse( |
|
choices=[{ |
|
"index": 0, |
|
"message": { |
|
"role": "assistant", |
|
"content": response['choices'][0]['text'].strip() |
|
}, |
|
"finish_reason": "stop" |
|
}], |
|
usage={ |
|
"prompt_tokens": len(prompt), |
|
"completion_tokens": len(response['choices'][0]['text']), |
|
"total_tokens": len(prompt) + len(response['choices'][0]['text']) |
|
} |
|
) |
|
except Exception as e: |
|
raise HTTPException(status_code=500, detail=str(e)) |
|
|
|
|
|
@app.get("/health") |
|
def health_check(): |
|
return {"status": "healthy"} |
|
|
|
|
|
if __name__ == "__main__": |
|
uvicorn.run(app, host="0.0.0.0", port=7860) |
