Upload 7 files

initial command

app.py

@@ -0,0 +1,22 @@
+try:
+  import gradio as gr
+except:
+  !pip install gradio
+  import gradio as gr
+from predict import predict_gradio
+
+title = "Mohammad Ali Food Classification🍔"
+description = "This demo is just for my training, Thank you Daniel Bourke for your courses"
+
+
+demo = gr.Interface(
+    predict_gradio,
+    inputs = gr.Image(type="pil"),
+    outputs = [gr.Label(num_top_classes = 3 , label = "All Predictions in All Classes"),
+     gr.Label(num_top_classes = 1 , label = "Model Predicts Image as a")],
+    examples=["/demo/examples/476421.jpg", "/demo/examples/3479599.jpg", "/demo/examples/911808.jpg"],
+    title = title,
+    description = description
+)
+
+demo.launch()

efficient_b2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e5872189fc92a388af899d85b3ccb27560cf1fa19330d3f363ee4f2a9ef8cc3
+size 31345215

model.py

@@ -0,0 +1,9 @@
+import torch
+import torchvision
+from torchvision.models import EfficientNet_B2_Weights , efficientnet_b2
+
+efficient_weight = EfficientNet_B2_Weights.DEFAULT
+
+efficient_transformer = efficient_weight.transforms()
+
+efficient_model = torch.load("/demo/efficient_b2.pth").to("cpu")

predict.py

@@ -0,0 +1,20 @@
+import torch
+import torchvision
+from model import efficient_transformer , efficient_model
+
+CLASS_NAMES = ['pizza', 'steak', 'sushi']
+
+def predict_gradio(image):
+  
+  image = efficient_transformer(image)
+
+  efficient_model.eval()
+
+  with torch.no_grad():
+    pred = efficient_model(torch.unsqueeze(image , dim = 0))
+
+  prediction_per_labels = {CLASS_NAMES[i]: float(torch.sigmoid(pred[0][i])) for i in range(len(CLASS_NAMES))}
+
+  prediction = CLASS_NAMES[torch.argmax(pred).item()]
+
+  return prediction_per_labels , prediction