Vision-language models (GPT-4V) in Prompt Engineering / GenAI - Model Metrics & Evaluation

Metrics & Evaluation - Vision-language models (GPT-4V)

Which metric matters for Vision-language models (GPT-4V) and WHY

Vision-language models like GPT-4V combine images and text. We want to check how well the model understands both. Key metrics include accuracy for classification tasks, BLEU or ROUGE for text generation quality, and precision and recall when detecting objects or answering questions about images. These metrics tell us if the model gives correct answers, describes images well, and finds important details without too many mistakes.

Confusion matrix example for image question answering

      | Predicted Yes | Predicted No |
      |---------------|--------------|
      | True Positives (TP) = 80  | False Negatives (FN) = 20 |
      | False Positives (FP) = 10 | True Negatives (TN) = 90 |

      Total samples = 80 + 20 + 10 + 90 = 200

      Precision = TP / (TP + FP) = 80 / (80 + 10) = 0.89
      Recall = TP / (TP + FN) = 80 / (80 + 20) = 0.80
      Accuracy = (TP + TN) / Total = (80 + 90) / 200 = 0.85
    

This matrix helps us see where the model makes mistakes: missing true answers (FN) or giving wrong answers (FP).

Precision vs Recall tradeoff with examples

Imagine GPT-4V is used to detect objects in images and describe them:

• High precision, low recall: The model only says "cat" when very sure, so it rarely makes mistakes (few false cats), but it misses some cats in images. Good if you want to avoid wrong labels.
• High recall, low precision: The model tries to find all cats, even if unsure, so it finds most cats but sometimes calls other animals cats by mistake. Good if missing any cat is bad.

Choosing depends on the task: for safety-critical tasks, recall is more important; for user experience, precision might matter more.

What "good" vs "bad" metric values look like for GPT-4V

Good metrics mean the model understands images and text well:

• Accuracy above 85% on image classification or question answering.
• Precision and recall both above 80%, showing balanced detection and correctness.
• BLEU or ROUGE scores above 0.5 for generated captions or answers, meaning text is relevant and fluent.

Bad metrics show problems:

• Accuracy below 60%, meaning many wrong answers.
• Precision very low (<50%) means many false positives.
• Recall very low (<50%) means many missed true cases.
• Very low BLEU/ROUGE (<0.3) means poor text quality.

Common pitfalls in evaluating vision-language models

• Accuracy paradox: High accuracy can be misleading if data is unbalanced (e.g., many easy images).
• Data leakage: If test images or captions appear in training, metrics look better but model won't generalize.
• Overfitting: Model performs well on training but poorly on new images, showing metrics that don't reflect real use.
• Ignoring metric tradeoffs: Focusing only on accuracy without precision/recall can hide important errors.
• Using wrong metrics: BLEU or ROUGE are for text quality, not classification accuracy.

Self-check question

Your GPT-4V model has 98% accuracy but only 12% recall on detecting rare objects in images. Is it good for production? Why or why not?

Answer: No, it is not good. The high accuracy is misleading because the rare objects are few, so the model mostly guesses "no" and is right. But the very low recall means it misses almost all rare objects, which is bad if detecting them is important. You need to improve recall to catch more true cases.

Key Result

Balanced precision and recall above 80% with good text quality scores indicate a well-performing vision-language model.

Practice

1. What is the main capability of vision-language models like GPT-4V?

easy

A. They understand and generate responses based on both images and text.

B. They only process text data without images.

C. They only analyze images without any text understanding.

D. They translate languages without any image input.

Solution

1. Step 1: Understand the model's input types

   Vision-language models take both images and text as input to understand context.

2. Step 2: Recognize the model's output capabilities

   They generate responses that relate to both the visual content and the text prompt.

3. Final Answer:

   They understand and generate responses based on both images and text. -> Option A

4. Quick Check:

   Vision + Language = Both inputs [OK]

Hint: Vision-language means both image and text understanding [OK]

Common Mistakes:

• Thinking the model only works with text
• Assuming it only processes images
• Confusing translation with vision-language tasks

2. Which of the following is the correct way to prompt GPT-4V to describe an image?

easy

A. Translate the text: [image]

B. Describe the image: [image]

C. Calculate the sum: [image]

D. Play music from: [image]

Solution

1. Step 1: Identify the prompt that asks for image description

   Only Describe the image: [image] clearly requests a description of the image content.

2. Step 2: Eliminate unrelated commands

   Options B, C, and D ask for translation, calculation, or music playing, which are unrelated to image description.

3. Final Answer:

   <code>Describe the image: [image]</code> -> Option B

4. Quick Check:

   Prompt matches task: describe image [OK]

Hint: Look for prompt asking to describe the image [OK]

Common Mistakes:

• Choosing prompts unrelated to images
• Confusing translation with description
• Ignoring the image context in the prompt

3. Given the following code snippet using GPT-4V API, what will be the output?

response = gpt4v.ask(image='cat.jpg', prompt='What animal is in the picture?')
print(response)

medium

A. SyntaxError: missing argument

B. "I cannot see any animal in the picture."

C. "The animal in the picture is a cat."

D. "The picture shows a dog."

Solution

1. Step 1: Understand the prompt and image input

   The prompt asks what animal is in the image named 'cat.jpg', which likely contains a cat.

2. Step 2: Predict the model's response

   GPT-4V will analyze the image and respond with the correct animal, which is a cat.

3. Final Answer:

   "The animal in the picture is a cat." -> Option C

4. Quick Check:

   Image name + prompt = cat answer [OK]

Hint: Match image content with prompt question [OK]

Common Mistakes:

• Assuming the model cannot see images
• Expecting error due to missing arguments
• Confusing animal types in output

4. Identify the error in this GPT-4V usage code snippet:

response = gpt4v.ask(prompt='Describe this image.')
print(response)

medium

A. Missing image input argument in the ask function.

B. The prompt text is too short.

C. The print statement is incorrect syntax.

D. The ask function does not exist in GPT-4V.

Solution

1. Step 1: Check required inputs for vision-language query

   GPT-4V requires both an image and a prompt to answer about the image.

2. Step 2: Identify missing argument

   The code only provides a prompt but no image, which is necessary for vision understanding.

3. Final Answer:

   Missing image input argument in the ask function. -> Option A

4. Quick Check:

   Image missing in ask() call [OK]

Hint: Vision queries need both image and prompt [OK]

Common Mistakes:

• Ignoring the need for image input
• Thinking prompt length causes error
• Assuming print syntax is wrong

5. You want GPT-4V to find all objects in a complex image and list them with counts. Which approach is best?

hard

A. Send multiple images without prompts and combine answers manually.

B. Send only the image without any prompt and expect a list.

C. Use a prompt asking to translate the image content to another language.

D. Use a prompt like List all objects and their counts in this image: [image] and parse the response.

Solution

1. Step 1: Understand the task requirements

   The task is to identify and count objects in one image, so a clear prompt is needed.

2. Step 2: Choose the prompt that requests object listing and counting

   Use a prompt like List all objects and their counts in this image: [image] and parse the response, which explicitly asks for listing objects and counts, which GPT-4V can handle.

3. Step 3: Eliminate other options

   Sending only the image without any prompt lacks specific task instructions. Using a prompt to translate the image content is unrelated to object detection. Sending multiple images without prompts and combining answers manually is inefficient and unclear.

4. Final Answer:

   Use a prompt like <code>List all objects and their counts in this image: [image]</code> and parse the response. -> Option D

5. Quick Check:

   Clear prompt + image = correct object list [OK]

Hint: Always include clear prompt with image for object detection [OK]

Common Mistakes:

• Sending image without prompt expecting detailed output
• Confusing translation with object detection
• Using multiple images without clear instructions