Model optimization (pruning, quantization) in Computer Vision - Model Metrics & Evaluation

Metrics & Evaluation - Model optimization (pruning, quantization)

Which metric matters for Model optimization (pruning, quantization) and WHY

When optimizing models by pruning or quantization, the key metrics to watch are accuracy or task-specific performance (like classification accuracy or mean average precision). This is because pruning and quantization reduce model size and speed up inference but can hurt accuracy if done too aggressively. We want to keep accuracy high while making the model smaller and faster.

Additionally, model size (memory footprint) and inference latency (speed) are important metrics to measure the benefits of optimization.

Confusion matrix example after pruning

    Original model confusion matrix:
      TP=90  FP=10
      FN=5   TN=95

    After pruning:
      TP=85  FP=15
      FN=10  TN=90

    Total samples = 200

    Precision before pruning = 90 / (90 + 10) = 0.9
    Recall before pruning = 90 / (90 + 5) = 0.947

    Precision after pruning = 85 / (85 + 15) = 0.85
    Recall after pruning = 85 / (85 + 10) = 0.895
    

This shows a slight drop in precision and recall after pruning, which is common if pruning is too aggressive.

Tradeoff: Accuracy vs Model Size and Speed

Pruning and quantization reduce model size and speed up predictions but can lower accuracy.

For example, a mobile app needs a small, fast model. It may accept a small accuracy drop to run smoothly on phones.

But a medical image model must keep very high accuracy, so pruning or quantization must be gentle or avoided.

Choosing the right balance depends on the use case: speed and size vs accuracy.

What "good" vs "bad" metric values look like for Model optimization

Good: Accuracy drops less than 1-2% after pruning or quantization, with model size reduced by 50% or more, and inference speed improved significantly.

Bad: Accuracy drops more than 5%, or the model becomes unstable, even if size and speed improve. This means the optimization hurt the model too much.

Common pitfalls in metrics for Model optimization

• Ignoring accuracy drop: Only measuring size and speed but missing that accuracy fell too much.
• Not testing on real data: Optimizing on training data can hide accuracy loss on new data.
• Over-pruning: Removing too many weights causes big accuracy loss.
• Quantization errors: Using too low precision can cause unstable predictions.

Self-check question

Your model optimization reduced size by 60% and sped up inference by 3x, but accuracy dropped from 95% to 85%. Is this good for production? Why or why not?

Answer: Usually no. A 10% accuracy drop is large and may hurt user experience or safety. The speed and size gains are good, but the accuracy loss is too high. You should try less aggressive pruning or quantization to keep accuracy higher.

Key Result

Model optimization aims to reduce size and speed up inference while keeping accuracy loss minimal (ideally under 2%).

Practice

1. What is the main goal of model pruning in computer vision?

easy

A. To remove less important parts of the model to reduce size

B. To increase the number of layers in the model

C. To add more training data for better accuracy

D. To convert the model to a different programming language

Solution

1. Step 1: Understand pruning concept

   Pruning means removing parts of the model that contribute less to its output.

2. Step 2: Identify pruning goal

   The goal is to reduce model size and speed up inference by cutting unnecessary parts.

3. Final Answer:

   To remove less important parts of the model to reduce size -> Option A

4. Quick Check:

   Pruning = Remove less important parts [OK]

Hint: Pruning cuts unneeded parts to shrink model size [OK]

Common Mistakes:

• Thinking pruning adds layers instead of removing
• Confusing pruning with data augmentation
• Believing pruning changes programming language

2. Which of the following is the correct way to apply quantization in TensorFlow Lite?

easy

A. model = tf.lite.TFLiteConverter.from_keras_model(model).convert()

B. converter.optimizations = [tf.lite.Optimize.DEFAULT]

C. model.compile(optimizer='adam', loss='mse')

D. model.fit(x_train, y_train, epochs=10)

Solution

1. Step 1: Identify quantization syntax

   In TensorFlow Lite, quantization is enabled by setting converter.optimizations to Optimize.DEFAULT.

2. Step 2: Check other options

   model = tf.lite.TFLiteConverter.from_keras_model(model).convert() converts model but does not enable quantization. Options B and C are training commands, not quantization.

3. Final Answer:

   converter.optimizations = [tf.lite.Optimize.DEFAULT] -> Option B

4. Quick Check:

   Quantization flag = converter.optimizations [OK]

Hint: Quantization needs converter.optimizations set to Optimize.DEFAULT [OK]

Common Mistakes:

• Confusing model conversion with quantization
• Using training commands instead of conversion flags
• Missing the optimization setting for quantization

3. Given this PyTorch pruning code snippet, what will be the output size of the model's first linear layer weights after pruning 20% of connections?

import torch
import torch.nn.utils.prune as prune

model = torch.nn.Sequential(
    torch.nn.Linear(100, 50),
    torch.nn.ReLU()
)
prune.l1_unstructured(model[0], name='weight', amount=0.2)
pruned_weights = model[0].weight
print((pruned_weights != 0).sum().item())

medium

A. 8000

B. 5000

C. 10000

D. 4000

Solution

1. Step 1: Calculate total weights

   The first linear layer has 100 inputs and 50 outputs, so total weights = 100 * 50 = 5000.

2. Step 2: Calculate remaining weights after pruning

   Pruning 20% removes 20% of weights, so remaining weights = 80% of 5000 = 4000.

3. Step 3: Understand pruning method

   PyTorch's l1_unstructured pruning does not remove weights but masks them, so the weight tensor size remains 5000, but the number of non-zero weights is 4000.

4. Step 4: Check print output

   The print statement counts non-zero weights, so output is 4000.

5. Final Answer:

   4000 -> Option D

6. Quick Check:

   5000 * 0.8 = 4000 [OK]

Hint: Remaining weights = total * (1 - pruning amount) [OK]

Common Mistakes:

• Calculating total weights incorrectly
• Using pruning amount as remaining instead of removed
• Confusing layer input/output dimensions

4. You tried to quantize a model but got an error: AttributeError: 'TFLiteConverter' object has no attribute 'optimizations'. What is the likely cause?

medium

A. Quantization requires training the model again

B. Model is too large to quantize

C. Using an outdated TensorFlow version without quantization support

D. The model has no weights to quantize

Solution

1. Step 1: Understand the error

   The error says the converter object lacks 'optimizations' attribute, meaning the TensorFlow version is old.

2. Step 2: Identify cause

   Older TensorFlow versions do not support the 'optimizations' attribute needed for quantization.

3. Final Answer:

   Using an outdated TensorFlow version without quantization support -> Option C

4. Quick Check:

   Missing attribute = outdated TensorFlow [OK]

Hint: Check TensorFlow version supports quantization features [OK]

Common Mistakes:

• Assuming model size causes attribute error
• Thinking quantization needs retraining always
• Believing model without weights causes this error

5. You want to deploy a computer vision model on a mobile device with limited memory and CPU. Which combination of optimization techniques is best to reduce model size and speed up inference without much accuracy loss?

hard

A. Apply pruning to remove unimportant weights, then quantize weights to 8-bit integers

B. Only increase model layers to improve accuracy

C. Use full precision weights and no pruning for best accuracy

D. Train longer without any model size changes

Solution

1. Step 1: Understand device constraints

   Mobile devices have limited memory and CPU, so model size and speed matter.

2. Step 2: Choose optimization techniques

   Pruning removes unnecessary weights reducing size; quantization reduces number precision speeding inference.

3. Step 3: Combine pruning and quantization

   Using both together reduces size and speeds up model with minimal accuracy loss.

4. Final Answer:

   Apply pruning to remove unimportant weights, then quantize weights to 8-bit integers -> Option A

5. Quick Check:

   Pruning + quantization = smaller, faster model [OK]

Hint: Combine pruning and quantization for efficient mobile models [OK]

Common Mistakes:

• Only increasing layers without optimization
• Ignoring quantization benefits
• Assuming full precision is always best for deployment