app.py

from fastapi import FastAPI, File, UploadFile
from fastapi.responses import JSONResponse
import tensorflow as tf
import numpy as np
from PIL import Image
import io

app = FastAPI()

model = tf.keras.models.load_model(
    "model_fixed.h5",
    compile=False
)


# Classes
CLASS_NAMES = [
    "Apple_Apple_Scab",
    "Apple_Black_Rot",
    "Apple_Cedar_Rust",
    "Apple_Healthy",
    "Blueberry_Healthy",
    "Cherry_Powdery_Mildew",
    "Cherry_Healthy",
    "Corn_Cercospora_Leaf_Spot",
    "Corn_Common_Rust",
    "Corn_Northern_Leaf_Blight",
    "Corn_Healthy",
    "Grape_Black_Rot",
    "Grape_Esca",
    "Grape_Leaf_Blight",
    "Grape_Healthy",
    "Peach_Bacterial_Spot",
    "Peach_Healthy",
    "Pepper_Bacterial_Spot",
    "Pepper_Healthy",
    "Potato_Early_Blight",
    "Potato_Late_Blight",
    "Potato_Healthy",
    "Strawberry_Leaf_Scorch",
    "Strawberry_Healthy",
    "Soybean_Healthy",
    "Squash_Powdery_Mildew",
    "Raspberry_Healthy",
    "Tomato_Bacterial_Spot",
    "Tomato_Early_Blight",
    "Tomato_Late_Blight",
    "Tomato_Leaf_Mold",
    "Tomato_Septoria_Leaf_Spot",
    "Tomato_Spider_Mites",
    "Tomato_Target_Spot",
    "Tomato_Yellow_Leaf_Curl_Virus",
    "Tomato_Mosaic_Virus",
    "Tomato_Healthy",
    "Orange_Citrus_Greening"
]

IMG_SIZE = 224

def predict(image):

    image = image.resize((IMG_SIZE, IMG_SIZE))

    image = np.array(image)

    image = tf.keras.applications.efficientnet.preprocess_input(image)

    image = np.expand_dims(image, axis=0)

    prediction = model.predict(image)[0]

    predicted_class = CLASS_NAMES[np.argmax(prediction)]

    confidence = float(np.max(prediction)) * 100

    return {
        "prediction": predicted_class,
        "confidence": f"{confidence:.2f}%"
    }

demo = gr.Interface(
    fn=predict,
    inputs=gr.Image(type="pil"),
    outputs="json",
    title="Plant Disease Classifier"
)

demo.launch()