Update app.py

app.py

@@ -1,18 +1,63 @@
+import torch
+from torch.utils.data import Dataset, DataLoader
+import pandas as pd
+import numpy as np
+import io
+import joblib
+import requests
+from tqdm import tqdm
+from PIL import Image
+from torchvision import transforms
+from sklearn.preprocessing import LabelEncoder
+from sklearn.model_selection import train_test_split
+from torchvision import models
 import gradio as gr
-from transformers import pipeline
 
-pipeline = pipeline(task="image-classification", model="julien-c/hotdog-not-hotdog")
+device = 'cpu'
+le = LabelEncoder()
+le = joblib.load("/kaggle/working/SVD/le.gz")
+
+class ModelPre(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.embedding = torch.nn.Sequential(
+            *list(models.convnext_small(weights=models.ConvNeXt_Small_Weights.IMAGENET1K_V1).children())[:-1],
+            torch.nn.Flatten(),
+            torch.nn.Linear(in_features=768, out_features=512),
+            torch.nn.ReLU(),
+            torch.nn.Linear(in_features=512, out_features=len(le.classes_) + 1),
+        )
+
+    def forward(self, data):
+        return self.embedding(data)
+
+model = torch.load("/SVD/GeoG.pth", map_location=torch.device(device))
+
+modelm = ModelPre()
+modelm.load_state_dict(model['model'])
+
+import warnings
+warnings.filterwarnings("ignore", category=RuntimeWarning, module="multiprocessing.popen_fork")
+
+cmp = transforms.Compose([
+    transforms.ToTensor(),
+    transforms.Resize(size=(224, 224), antialias=True),
+    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
+])
 
 def predict(input_img):
-    predictions = pipeline(input_img)
-    return input_img, {p["label"]: p["score"] for p in predictions} 
+    with torch.inference_mode():
+        img = cmp(input_img).unsqueeze(0)
+        res = modelm(img.to(device))
+        prediction = le.inverse_transform(torch.argmax(res.cpu()).unsqueeze(0).numpy())[0]
+        return prediction
 
 gradio_app = gr.Interface(
-    predict,
-    inputs=gr.Image(label="Select hot dog candidate", sources=['upload', 'webcam'], type="pil"),
-    outputs=[gr.Image(label="Processed Image"), gr.Label(label="Result", num_top_classes=2)],
-    title="Hot Dog? Or Not?",
+    fn=predict,
+    inputs=gr.Image(label="Upload an Image", type="pil"),
+    outputs=gr.Label(label="Location"),
+    title="Predict the Location of this Image"
 )
 
 if __name__ == "__main__":
-    gradio_app.launch()
+    gradio_app.launch()