Upload app.py with huggingface_hub

app.py

@@ -0,0 +1,75 @@
+import gradio as gr
+import torch
+import torch.nn.functional as F
+import numpy as np
+from torchvision import models, transforms
+from huggingface_hub import hf_hub_download
+from PIL import Image
+from pytorch_grad_cam import GradCAM
+from pytorch_grad_cam.utils.image import show_cam_on_image
+
+MODEL_REPO = "ummanmm/classroom-engagement-mobilenet"
+CLASSES = [
+    "Oriented (Lecture-Focused)",
+    "Diverted (Looking Away/Down)",
+    "Obscured (Blocked/Empty)",
+]
+IMAGENET_MEAN = [0.485, 0.456, 0.406]
+IMAGENET_STD = [0.229, 0.224, 0.225]
+
+model_path = hf_hub_download(
+    repo_id=MODEL_REPO, filename="best_baseline.pth"
+)
+
+model = models.mobilenet_v2(weights=None)
+model.classifier[1] = torch.nn.Linear(model.last_channel, 3)
+model.load_state_dict(
+    torch.load(model_path, map_location=torch.device("cpu"))
+)
+model.eval()
+
+transform = transforms.Compose([
+    transforms.Resize((224, 224)),
+    transforms.ToTensor(),
+    transforms.Normalize(mean=IMAGENET_MEAN, std=IMAGENET_STD),
+])
+
+
+def predict_engagement(image):
+    if image is None:
+        return {}, None
+
+    img = Image.fromarray(image).convert("RGB")
+    tensor = transform(img).unsqueeze(0)
+
+    with torch.no_grad():
+        outputs = model(tensor)
+        probs = F.softmax(outputs[0], dim=0)
+
+    confidences = {CLASSES[i]: float(probs[i]) for i in range(3)}
+
+    rgb = np.array(img.resize((224, 224))).astype(np.float32) / 255.0
+    cam = GradCAM(model=model, target_layers=[model.features[-1]])
+    grayscale = cam(input_tensor=tensor, targets=None)[0, :]
+    heatmap = show_cam_on_image(rgb, grayscale, use_rgb=True)
+
+    return confidences, heatmap
+
+
+demo = gr.Interface(
+    fn=predict_engagement,
+    inputs=gr.Image(label="Upload Student Crop"),
+    outputs=[
+        gr.Label(num_top_classes=3, label="Engagement Prediction"),
+        gr.Image(label="Grad-CAM Attention Heatmap"),
+    ],
+    title="Real-Time Classroom Engagement Telemetry",
+    description=(
+        "Upload a cropped image of a student to classify their "
+        "physical engagement state using a custom MobileNetV2 model. "
+        "The Grad-CAM heatmap shows which regions the CNN focuses on."
+    ),
+    examples=[],
+)
+
+demo.launch()