How Should You Train a Traffic Sign Classifier on Vertex AI?
Short answer
D Develop the model training code for image classification, and train a model by using Vertex AI custom training.
Question
Key Takeaways
The exam tests your ability to distinguish image classification from object detection and to identify when custom training is required for tuning. The trap is that 'identify traffic signs' sounds like detection, but each image is already cropped to a single sign, making classification the appropriate task.
The correct approach is to develop custom image classification training code with Vertex AI, not AutoML or object detection, because the dataset contains cropped images of one sign per image. Community consensus strongly favors option D (77%), with the key clue being the requirement to tune the model during each training run.
Common mistake: The most common wrong answer is C (custom object detection). Test takers over-interpret 'identify' as detection, but the images are already cropped to show exactly one traffic sign, so bounding-box detection is unnecessary and classification is sufficient.
Expert Analysis
Why the Answer Is Correct
Option D correctly matches the problem requirements: images contain one cropped traffic sign, so it is a single-label image classification task. Because the question states you need to 'tune the model during each training run,' AutoML options (A and B) are not suitable—AutoML hides training details and limits direct tuning flexibility. Custom training (C or D) allows you to modify hyperparameters, architecture, and training loops. Between C and D, classification is the right problem type because no bounding boxes are needed when each image already shows exactly one sign.
Why the Other Options Are Wrong
Options A and B fail because AutoML does not provide the tuning flexibility explicitly mentioned in the question. Option C fails because it uses custom training but incorrectly frames the task as object detection; the images are already cropped to a single traffic sign, so returning bounding boxes is overkill. Community comments [1] and [6] correctly note that object detection is unnecessary because the images are already nicely cropped. Comments [2] and [5] argue for C by treating 'identify' as detection, but the 'cropped to one out of ten' clue is a clear signal for classification.
Community Comment Notes
Comment [1], with 16 likes, correctly explains that AutoML lacks tuning flexibility and that object detection isn't needed since images are cropped to a single sign. Comment [3] highlights the ambiguity between classification and detection, but ultimately the 'cropped to a single traffic light' detail points to D. Comment [6] reinforces that object detection would be overkill if the images are already cleanly cropped. A minority (23%) chose C, but their reasoning generally depends on a broader interpretation of 'identify' rather than the specific dataset structure given in the question.
Official Reference
https://cloud.google.com/vertex-ai/docs/training/custom-training
This link opens the official vendor documentation in a new tab.
Exam Strategy
Look for key phrases like 'cropped to show one out of ten' to determine classification versus object detection. When the question mentions tuning the model during each training run, reject AutoML and choose custom training; if images are already centered on a single object, choose image classification over object detection.
Community Votes
77% of anonymous learners picked answer D. Votes are pick records left by other test-takers — they are not the verified answer.
Discussion 7 contributions
Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.
From public community threads 7
-
pikachu007 👍 16 Answered DNot A or B since automl doesnt provide you without flexibility to tune. Not C because object detection is not required since the images are cropped to a single traffic light
-
guilhermebutzke 👍 3 Answered CCorrect: C The phrases "identify traffic signs in images extracted from videos" and "images that were cropped to show one out of ten different traffic signs" suggest that this is an image detection problem. The first phrase appears to have the same meaning as "images with," and the second phrase suggests that only one type of traffic sign was used in the problem, indicating that it cannot be used in a multi-class problem. For all these reasons, I believe the best option is C.
-
baimus 👍 2 Answered CThere is literally no way to know if this is C or D, as "labelled" and "identify street signs" are too ambiguous to know if its detection or classification. The "cropped to a single traffic light" seems like maybe D, but that's hardly ML knowledge, it's a pub quiz guess.
-
Dirtie_Sinkie 👍 1 Answered DAnswer is D, not C in my opinion. Object detection might be used in a real-world project since there are a lot of variables which may affect the picture like visibility, colour fading, having other things in the picture like streetlights, birds, etc. Way too overkill for our question here. I'm assuming the pictures are already nicely cropped out with none of this extra stuff in the pictures.
-
kornick 👍 1 Answered DNot C because object detection is not required since the images are cropped to a single traffic light
Show 2 more comments
-
shaoshao 👍 1 Answered CC. "identify" a small signal is an object detection task, not a classification task.
-
gscharly 👍 1 Answered Dagree with pikachu007
Reader comments & corrections
No reader corrections yet — spotted an error or have a note? Share it below.
Related Analysis
Practice All PMLE Questions
Access 65 questions with complete answers and detailed explanations.
View Full PMLE Practice Test →