appImage.py

"""from fastapi import FastAPI, UploadFile, File
from fastapi.responses import RedirectResponse, JSONResponse
from transformers import AutoProcessor, AutoModelForCausalLM
from PIL import Image
import tempfile
import torch

app = FastAPI()

# Load model
try:
    processor = AutoProcessor.from_pretrained("microsoft/git-large-coco")
    model = AutoModelForCausalLM.from_pretrained("microsoft/git-large-coco")
    USE_GIT = True
except Exception:
    from transformers import pipeline
    captioner = pipeline("image-to-text", model="nlpconnect/vit-gpt2-image-captioning")
    USE_GIT = False

def generate_caption(image_path):
    try:
        if USE_GIT:
            image = Image.open(image_path)
            inputs = processor(images=image, return_tensors="pt")
            outputs = model.generate(**inputs, max_length=50)
            return processor.batch_decode(outputs, skip_special_tokens=True)[0]
        else:
            result = captioner(image_path)
            return result[0]['generated_text']
    except Exception as e:
        return f"Error generating caption: {str(e)}"

@app.post("/imagecaption/")
async def caption_from_frontend(file: UploadFile = File(...)):
    contents = await file.read()
    with tempfile.NamedTemporaryFile(delete=False, suffix=".png") as tmp:
        tmp.write(contents)
        image_path = tmp.name

    caption = generate_caption(image_path)
    return JSONResponse({"caption": caption})

@app.get("/")
def home():
    return RedirectResponse(url="/")"""
# appImage.py
from transformers import pipeline, AutoProcessor, AutoModelForCausalLM
import tempfile, os
from PIL import Image
from gtts import gTTS
import torch

try:
    processor = AutoProcessor.from_pretrained("microsoft/git-large-coco")
    model = AutoModelForCausalLM.from_pretrained("microsoft/git-large-coco")
    USE_GIT = True
except Exception:
    captioner = pipeline("image-to-text", model="nlpconnect/vit-gpt2-image-captioning")
    USE_GIT = False

async def caption_image(file):
    contents = await file.read()
    with tempfile.NamedTemporaryFile(delete=False, suffix=".png") as tmp:
        tmp.write(contents)
        image_path = tmp.name

    if USE_GIT:
        image = Image.open(image_path).convert('RGB')
        pixel_values = processor(images=image, return_tensors="pt").pixel_values
        generated_ids = model.generate(pixel_values, max_length=500)
        caption = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
    else:
        captions = captioner(image_path)
        caption = captions[0]['generated_text'] if captions else "No caption generated."

    audio_path = text_to_speech(caption)

    result = {"caption": caption}
    if audio_path:
        result["audio"] = f"/files/{os.path.basename(audio_path)}"
    return result

def text_to_speech(text: str):
    try:
        tts = gTTS(text)
        temp_audio = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
        tts.save(temp_audio.name)
        return temp_audio.name
    except:
        return ""