File size: 2,562 Bytes
245c6b4
1f97cf6
 
 
 
 
9eec479
 
 
1f97cf6
 
 
 
 
 
9eec479
 
 
245c6b4
1f97cf6
 
 
 
 
 
 
b15e27e
1f97cf6
 
 
 
 
 
 
 
 
 
245c6b4
 
1f97cf6
 
245c6b4
 
 
 
 
1f97cf6
245c6b4
 
 
 
1f97cf6
245c6b4
1f97cf6
 
245c6b4
1f97cf6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
245c6b4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
from fastapi import FastAPI, File, UploadFile, Form
from transformers import AutoModel, AutoTokenizer
import uvicorn
from PIL import Image
import io
import torch
import os
from huggingface_hub import login

app = FastAPI(title="Image-Text API")

# Model ve tokenizer için global değişkenler
model = None
tokenizer = None

# HuggingFace token ile giriş yap
if "HUGGINGFACE_TOKEN" in os.environ:
    login(token=os.environ["HUGGINGFACE_TOKEN"])

async def init_model():
    global model, tokenizer
    
    # Model ve tokenizer'ı yükle
    model = AutoModel.from_pretrained(
        'openbmb/MiniCPM-V-2_6',
        trust_remote_code=True,
        attn_implementation='eager',
        torch_dtype=torch.bfloat16
    )
    model = model.eval()
    if torch.cuda.is_available():
        model = model.cuda()
    
    tokenizer = AutoTokenizer.from_pretrained(
        'openbmb/MiniCPM-V-2_6',
        trust_remote_code=True
    )

@app.on_event("startup")
async def startup_event():
    await init_model()

@app.post("/process")
async def process_image_text(
    image: UploadFile = File(...),
    prompt: str = Form(...),
    stream: bool = Form(False)
):
    try:
        # Resmi oku ve PIL Image'a dönüştür
        image_content = await image.read()
        pil_image = Image.open(io.BytesIO(image_content)).convert('RGB')
        
        # Mesaj formatını hazırla
        msgs = [{'role': 'user', 'content': [pil_image, prompt]}]
        
        if stream:
            # Streaming yanıt için generator
            async def generate():
                result = model.chat(
                    image=None,
                    msgs=msgs,
                    tokenizer=tokenizer,
                    sampling=True,
                    stream=True
                )
                for text in result:
                    yield {"text": text}
            
            return generate()
        else:
            # Normal yanıt
            result = model.chat(
                image=None,
                msgs=msgs,
                tokenizer=tokenizer
            )
            
            return {
                "status": "success",
                "result": result
            }
            
    except Exception as e:
        return {
            "status": "error",
            "message": str(e)
        }

@app.get("/")
async def root():
    return {
        "message": "Image-Text API'ye hoş geldiniz",
        "usage": "POST /process endpoint'ini kullanın"
    }

if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=8000)