Upload 17 files

app.py

@@ -0,0 +1,61 @@
+### 1. Imports and class names setup ### 
+import gradio as gr
+import os
+import torch
+
+from model import create_effnetb2_model
+from timeit import default_timer as timer
+
+# Setup class names
+with open("class_names.txt", 'r') as f:
+  classes = [name.strip() for name in f]
+
+# Model and transforms
+model, transform = create_effnetb2_model(
+    num_classes=len(classes)
+)
+
+model.load_state_dict(
+    torch.load(
+        f="model_v3.pth",
+        map_location=torch.device("cpu")
+    )
+)
+
+# Predict function
+def predict(img):
+ 
+    start_time = timer()
+    
+    # Transform the target image and add a batch dimension
+    img = transform(img).unsqueeze(0)
+    
+    model.eval()
+    with torch.inference_mode():
+        
+        predictions = torch.softmax(model(img), dim=1)
+    
+    # Create a prediction label and prediction probability dictionary for each prediction class (this is the required format for Gradio)
+    pred_labels_and_probs = {classes[i]: float(predictions[0][i]) for i in range(len(classes))}
+    
+    pred_time = round(timer() - start_time, 4)
+    
+    return pred_labels_and_probs, pred_time
+
+example_list = [["examples/" + example] for example in os.listdir("examples")]
+
+# Gradio interface
+title = "Weather image recognition ⛅❄☔"
+description = "Classifies the weather from an image, able to recognize 12 types of weather."
+
+demo = gr.Interface(fn=predict,
+                    inputs=gr.Image(type="pil"),
+                    outputs=[gr.Label(num_top_classes=1, label="Predictions"), 
+                             gr.Number(label="Prediction time (s)")],
+                    examples=example_list,
+                    title=title,
+                    description=description)
+
+
+demo.launch(debug=False,
+            share=False)

class_names.txt

@@ -0,0 +1,12 @@
+cloudy
+dew
+fogsmog
+frost
+hail
+lightning
+rain
+rainbow
+shine
+snow
+sunrise
+tornado

model.py

@@ -0,0 +1,25 @@
+
+import torch
+import torchvision
+from torch import nn
+
+def create_effnetb2_model(num_classes: int):
+  
+    """Creates an EfficientNetB2 model."""
+   
+    # Create model and transforms
+    weights = torchvision.models.EfficientNet_B2_Weights.DEFAULT
+    transforms = weights.transforms()
+    model = torchvision.models.efficientnet_b2(weights=weights)
+
+    # Freeze layers
+    for param in model.parameters():
+        param.requires_grad = False
+
+    # Change classifier
+    model.classifier = nn.Sequential(
+        nn.Dropout(p=0.3, inplace=True),
+        nn.Linear(in_features=1408, out_features=num_classes)
+    )
+    
+    return model, transforms

model_v3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9dc5444d4c7cb46fc682a72c83f5cc3316240497e2b5e19f86186c0c866d6791
+size 31323721

requirements.txt

@@ -0,0 +1,3 @@
+torch==1.12.0
+torchvision==0.13.0
+gradio==3.1.4