add files

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+vit_model.pth filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,75 @@
+### 1. Imports and class names setup ### 
+import gradio as gr
+import os
+import torch
+
+from model import create_vit_model
+from timeit import default_timer as timer
+from typing import Tuple, Dict
+
+# Setup class names
+class_names = ['أ','ب','ت','ث','ج','ح','خ','د','ذ','ر','ز','س','ش','ص','ض','ط','ظ','ع','غ','ف','ق','ك','ل','م','ن','ه','و','ي']
+
+### Model and transforms preparation ###
+
+# Create ViTB16 model
+vit, vit_transforms = create_vit_model(
+     len(class_names)
+)
+
+# Load saved weights
+vit.load_state_dict(
+    torch.load(
+        f="vit_model.pth",
+        map_location=torch.device("cpu"),  # load to CPU
+    )
+)
+
+### Predict function ###
+
+# Create predict function
+def predict(img) -> Tuple[Dict, float]:
+    """Transforms and performs a prediction on img and returns prediction and time taken.
+    """
+    # Start the timer
+    start_time = timer()
+    
+    # Transform the target image and add a batch dimension
+    img = vit_transforms(img).unsqueeze(0)
+    
+    # Put model into evaluation mode and turn on inference mode
+    vit.eval()
+    with torch.inference_mode():
+        # Pass the transformed image through the model and turn the prediction logits into prediction probabilities
+        pred_probs = torch.softmax(vit(img), dim=1)
+    
+    # Create a prediction label and prediction probability dictionary for each prediction class (this is the required format for Gradio's output parameter)
+    pred_labels_and_probs = {class_names[i]: float(pred_probs[0][i]) for i in range(len(class_names))}
+    
+    # Calculate the prediction time
+    pred_time = round(timer() - start_time, 5)
+    
+    # Return the prediction dictionary and prediction time 
+    return pred_labels_and_probs, pred_time
+
+### Gradio app ###
+
+# Create title, description and article strings
+title = "Arabic Handwritten Characters Recognition"
+description = "A ViTB16 feature extractor computer vision model to classify hand written Arabic letters."
+
+# Create examples list from "examples/" directory
+example_list = [["examples/" + example] for example in os.listdir("examples")]
+
+# Create the Gradio demo
+demo = gr.Interface(fn=predict, # mapping function from input to output
+                    inputs=gr.Image(type="pil"), # what are the inputs?
+                    outputs=[gr.Label(num_top_classes=3, label="Predictions"), # what are the outputs?
+                             gr.Number(label="Prediction time (s)")], # our fn has two outputs, therefore we have two outputs
+                    # Create examples list from "examples/" directory
+                    examples=example_list, 
+                    title=title,
+                    description=description)
+
+# Launch the demo!
+demo.launch()

model.py

@@ -0,0 +1,42 @@
+import torch
+import torchvision
+
+from torch import nn
+
+def create_vit_model(num_classes:int=28, 
+                          seed:int=42):
+    """Creates an ViTB16 feature extractor model and transforms.
+
+    Args:
+        num_classes (int, optional): number of classes in the classifier head. 
+            Defaults to 28.
+        seed (int, optional): random seed value. Defaults to 42.
+
+    Returns:
+        model (torch.nn.Module): ViTB16 feature extractor model. 
+        transforms (torchvision.transforms): ViTB16 image transforms.
+    """
+    # Create ViTB16 pretrained weights, transforms and model
+    weights = torchvision.models.ViT_B_16_Weights.DEFAULT
+
+    # Get transforms from weights
+    vit_transforms = weights.transforms()
+
+    # Extend the vit_transforms to include grayscale conversion, since vit is trained on 3-channel RGB
+    transforms = transforms.Compose([
+        transforms.Grayscale(num_output_channels=3),  # Convert grayscale to 3-channel RGB
+        vit_transforms  # Append the existing transforms
+    ])
+    # transforms = weights.transforms()
+    model = torchvision.models.vit_b_16(weights=weights)
+
+    # Freeze all layers in base model
+    for param in model.parameters():
+        param.requires_grad = False
+
+    # Change heads with random seed for reproducibility
+    model.heads = torch.nn.Sequential(
+    nn.Linear(in_features=768,
+              out_features=28, # Number of Arabic letters = our classes
+              bias=True).to(device))
+    return model, transforms

requirements.txt

@@ -0,0 +1,3 @@
+torch
+torchvision
+gradio

vit_model.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8cd7bf531b4719876c51543e83e7c782b3eeeaf1f2cadbf1110ecc7e57308eee
+size 343341370