Vision Transformer (ViT) in Computer Vision - Model Metrics & Evaluation

Metrics & Evaluation - Vision Transformer (ViT)

Which metric matters for Vision Transformer (ViT) and WHY

For Vision Transformers used in image classification, accuracy is the main metric. It tells us how many images the model labels correctly out of all images. However, when classes are uneven or some mistakes cost more, precision, recall, and F1 score become important. These metrics help us understand if the model is good at finding certain classes or avoiding wrong guesses.

Confusion Matrix Example

      | Predicted Cat | Predicted Dog |
      |--------------|---------------|
      | True Cat: 45 | False Dog: 5  |
      | False Cat: 3 | True Dog: 47  |

      Total samples = 45 + 5 + 3 + 47 = 100

      Precision (Cat) = TP / (TP + FP) = 45 / (45 + 3) = 0.9375
      Recall (Cat) = TP / (TP + FN) = 45 / (45 + 5) = 0.9
    

Precision vs Recall Tradeoff with Examples

Imagine ViT is used to detect rare animals in photos. If we want to be sure when the model says "animal found," we need high precision. This means fewer false alarms. But if missing any rare animal is bad, we want high recall to catch as many as possible, even if some guesses are wrong.

Choosing between precision and recall depends on the task. For example, in medical image analysis, missing a disease (low recall) is worse than a false alarm (low precision). For general object recognition, balanced metrics like F1 score help.

Good vs Bad Metric Values for ViT

Good: Accuracy above 85% on a balanced dataset means the ViT is learning well. Precision and recall above 80% show it finds and labels classes reliably.

Bad: Accuracy near random chance (e.g., 10% for 10 classes) means the model is not learning. Very high accuracy but low recall means it misses many true cases. Low precision means many wrong guesses.

Common Pitfalls in Metrics for ViT

• Accuracy paradox: High accuracy can hide poor performance if classes are imbalanced.
• Data leakage: If test images are too similar to training, metrics look better but model won't generalize.
• Overfitting: Very high training accuracy but low test accuracy means the model memorizes training images, not learning patterns.

Self-Check Question

Your ViT model has 98% accuracy but only 12% recall on a rare class. Is it good for production? Why or why not?

Answer: No, it is not good. The model misses most true cases of the rare class (low recall), which can be critical depending on the task. High accuracy is misleading because the rare class is small compared to others.

Key Result

Accuracy is key for ViT image classification, but precision and recall reveal deeper performance, especially on rare or important classes.

Practice

1. What is the main purpose of splitting an image into patches in a Vision Transformer (ViT)?

easy

A. To reduce the image size by cropping

B. To convert the image into smaller parts that the transformer can process as tokens

C. To apply convolution filters on each patch separately

D. To increase the image resolution for better detail

Solution

1. Step 1: Understand ViT input processing

   ViT splits images into fixed-size patches to treat each patch like a word token in language models.

2. Step 2: Purpose of patch splitting

   This allows the transformer to process image patches as a sequence, enabling attention mechanisms to learn relationships.

3. Final Answer:

   To convert the image into smaller parts that the transformer can process as tokens -> Option B

4. Quick Check:

   Image patches = tokens for transformer [OK]

Hint: Think of patches as words in a sentence for the transformer [OK]

Common Mistakes:

• Confusing patch splitting with image resizing
• Thinking patches are processed by convolution
• Assuming patches increase image resolution

2. Which of the following is the correct way to add a class token to the patch embeddings in ViT using Python-like pseudocode?

easy

A. patches = torch.cat([class_token, patches], dim=1)

B. patches = torch.cat([patches, class_token], dim=1)

C. patches = torch.cat([patches, class_token], dim=0)

D. patches = torch.cat([class_token, patches], dim=0)

Solution

1. Step 1: Understand tensor concatenation dimension

   Patch embeddings are sequences along dimension 1 (batch, seq, embed); class token must be prepended along this dimension.

2. Step 2: Correct concatenation syntax

   Using torch.cat with dim=1 adds class_token at the start of the sequence correctly.

3. Final Answer:

   patches = torch.cat([class_token, patches], dim=1) -> Option A

4. Quick Check:

   Class token prepended along sequence dim = patches = torch.cat([class_token, patches], dim=1) [OK]

Hint: Class token goes first, concat along sequence dimension (dim=1) [OK]

Common Mistakes:

• Concatenating along wrong dimension (dim=0)
• Appending class token at the end instead of start
• Mixing order of tensors in concat

3. Given the following simplified ViT patch embedding code, what is the shape of patch_embeddings after processing a batch of 8 images of size 32x32 with patch size 8 and embedding dimension 64?

patch_size = 8
embedding_dim = 64
batch_size = 8
image_size = 32
num_patches = (image_size // patch_size) ** 2
patch_embeddings = torch.randn(batch_size, num_patches, embedding_dim)

medium

A. (16, 8, 64)

B. (8, 64, 16)

C. (8, 8, 64)

D. (8, 16, 64)

Solution

1. Step 1: Calculate number of patches

   Number of patches = (32 / 8)^2 = 4^2 = 16 patches per image.

2. Step 2: Determine patch_embeddings shape

   Shape is (batch_size, num_patches, embedding_dim) = (8, 16, 64).

3. Final Answer:

   (8, 16, 64) -> Option D

4. Quick Check:

   Batch=8, patches=16, embed=64 [OK]

Hint: Calculate patches as (image/patch)^2, then batch x patches x embed [OK]

Common Mistakes:

• Mixing embedding dimension and patch count order
• Calculating patches incorrectly
• Confusing batch size with patch count

4. You have this ViT code snippet that throws an error:

class_token = torch.randn(1, 1, 64)
patches = torch.randn(8, 16, 64)
input_seq = torch.cat([class_token, patches], dim=1)

What is the cause of the error?

medium

A. Embedding dimensions do not match

B. Wrong concatenation dimension; should be dim=0

C. class_token shape should be (8, 1, 64) to match batch size

D. Dimension mismatch because class_token sequence size is 1 but patches sequence size is 16

Solution

1. Step 1: Check batch size compatibility

   class_token has batch size 1, patches have batch size 8; they must match for concatenation.

2. Step 2: Fix class_token shape

   class_token should be repeated or created with shape (8, 1, 64) to match patches batch size.

3. Final Answer:

   class_token shape should be (8, 1, 64) to match batch size -> Option C

4. Quick Check:

   Batch sizes must match for concat [OK]

Hint: Match batch sizes before concatenating tensors [OK]

Common Mistakes:

• Ignoring batch size mismatch
• Changing wrong concat dimension
• Assuming embedding dims cause error

5. In a Vision Transformer model, why is the class token important for image classification tasks?

hard

A. It aggregates information from all patches via attention to produce a final image representation

B. It stores the positional information of patches

C. It applies convolution to patches before transformer layers

D. It normalizes the patch embeddings before feeding to the transformer

Solution

1. Step 1: Understand class token role

   The class token is a special token that attends to all patch tokens and gathers their information.

2. Step 2: Use in classification

   After transformer layers, the class token embedding is used as the image's summary representation for classification.

3. Final Answer:

   It aggregates information from all patches via attention to produce a final image representation -> Option A

4. Quick Check:

   Class token = image summary for classification [OK]

Hint: Class token collects info from patches for final decision [OK]

Common Mistakes:

• Confusing class token with positional encoding
• Thinking class token applies convolution
• Assuming class token normalizes embeddings