Deal of The Day! Hurry Up, Grab the Special Discount - Save 25% - Ends In 00:00:00 Coupon code: SAVE25
Welcome to Pass4Success

- Free Preparation Discussions
Mail Us support@pass4success.com
Location US
Mob_nav_barsMENU

Google Professional Machine Learning Engineer Exam - Topic 1 Question 19 Discussion

You have been asked to develop an input pipeline for an ML training model that processes images from disparate sources at a low latency. You discover that your input data does not fit in memory. How should you create a dataset following Google-recommended best practices?
B) Convert the images to tf .Tensor Objects, and then run Dataset. from_tensor_slices{).
A) Create a tf.data.Dataset.prefetch transformation
C) Convert the images to tf .Tensor Objects, and then run tf. data. Dataset. from_tensors ().
D) Convert the images Into TFRecords, store the images in Cloud Storage, and then use the tf. data API to read the images for training

Google Professional Machine Learning Engineer Exam - Topic 1 Question 19 Discussion

Actual exam question for Google's Professional Machine Learning Engineer exam
Question #: 19
Topic #: 1
[All Professional Machine Learning Engineer Questions]

You have been asked to develop an input pipeline for an ML training model that processes images from disparate sources at a low latency. You discover that your input data does not fit in memory. How should you create a dataset following Google-recommended best practices?

Show Suggested Answer Hide Answer
Suggested Answer: B

by Haydee at May 09, 2022, 08:56 AM

Limited Time Offer

25%

Off

Get Premium Professional Machine Learning Engineer Questions as Interactive Web-Based Practice Test or PDF

Contribute your Thoughts:

0/2000 characters
Submit Cancel
Karan
9 months ago
D is the best practice according to Google, no doubt!
upvoted 0 times
...
Lou
9 months ago
B sounds tempting, but it won't handle the memory issue well.
upvoted 0 times
...
Chaya
9 months ago
Wait, are we really using Cloud Storage for this? Seems like overkill.
upvoted 0 times
...
Nan
9 months ago
I think A is a good option too, but not the best for this scenario.
upvoted 0 times
...
Filiberto
9 months ago
D is definitely the way to go for large datasets!
upvoted 0 times
...
Lenna
10 months ago
I practiced a similar question, and I think using Cloud Storage with TFRecords is the most efficient way to handle large datasets, which points to D.
upvoted 0 times
...
Yolando
10 months ago
I think converting images to tensors is important, but I can't recall if it's better to use from_tensor_slices or from_tensors.
upvoted 0 times
...
Isadora
10 months ago
I remember we discussed using TFRecords for large datasets, so I think option D might be the right choice.
upvoted 0 times
...
Mohammad
10 months ago
I'm not entirely sure, but I feel like prefetching could help with latency, so maybe A is worth considering too?
upvoted 0 times
...
Malcolm
10 months ago
Okay, I've got it. The export policy on the local BGP peer is the way to go to reject the VPN routes before they get sent out.
upvoted 0 times
...
Jacob
10 months ago
Okay, I've got this. ARP poisoning is the MiTM attack mentioned in the question. Now I just need to make sure I explain my reasoning clearly in the exam.
upvoted 0 times
...
Cassandra
10 months ago
This looks like a tricky one. I'll need to carefully review the definitions of type 1 and type 2 hypervisors to identify the correct options.
upvoted 0 times
...
Krissy
10 months ago
I feel like I've seen similar questions before, and I want to say both strategies could possibly work, but I need to double-check my notes on swap exposures.
upvoted 0 times
...
Inocencia
10 months ago
Hmm, I'm a bit unsure about this one. The options seem similar, so I'll need to think it through carefully before selecting an answer.
upvoted 0 times
...

Save Cancel