MENU
A company wants to improve the sustainability of its ML operations.
Which actions will reduce the energy usage and computational resources that are associated with the company's training jobs? (Choose two.)
Suggested Answer: A, D
Get Premium MLA-C01 Questions
as Interactive Web-Based Practice Test or PDF
Real Exam Format
Detailed Explanations
Instant Access
Get Your Practice Test
28 comments · newest first
TY
Tyra
10 months ago
I’m not sure about D, are AWS Trainium instances really that efficient?
WE
Weldon
10 months ago
Totally agree, A makes a lot of sense!
DE
Denae
10 months ago
A is a solid choice for stopping wasted training time.
KA
Kattie
11 months ago
C could help, but I think A is the best option here!
MI
Ming
11 months ago
B seems irrelevant for energy savings, just saying.
YU
Yuonne
11 months ago
I recall that distributed training with PyTorch or TensorFlow can sometimes lead to higher resource consumption, so I'm not sure if that's the best choice for sustainability.
MA
Margart
11 months ago
I practiced a question similar to this, and I feel like deploying with AWS Lambda could help reduce resource usage, but I'm not confident.
SH
Shonda
11 months ago
I'm not entirely sure, but I think using AWS Trainium instances might be more efficient for training compared to regular instances.
DO
Domingo
12 months ago
I remember that using SageMaker Debugger can help stop jobs that aren't converging, which should save energy.
CA
Carin
12 months ago
The distributed training option in PyTorch or TensorFlow seems like it could be a good strategy to explore. I'll make sure to consider that as well.
KA
Karrie
12 months ago
I'm not familiar with some of these AWS services, so I'll need to do a quick review to make sure I understand how they can help with sustainability.
SO
Sophia
1 year ago
Using Amazon SageMaker Debugger to stop non-converging training jobs sounds like a good way to save resources. And using AWS Trainium instances could also be helpful.
ME
Merrilee
1 year ago
Hmm, I'm a bit unsure about this one. I'll need to think through the different options and how they might impact the company's sustainability goals.
AN
Annice
1 year ago
This question seems straightforward. I'll focus on options that can reduce energy usage and computational resources for the training jobs.
ED
Edna
1 year ago
B? Really? Ground Truth for data labeling? I guess they want us to label our data with crayons and glitter to be more 'sustainable'.
LU
Lura
1 year ago
D all the way! AWS Trainium instances are specifically designed for training, so they're bound to be more efficient.
KI
Kimbery
1 year ago
C: Deploying models by using AWS Lambda functions could also help in reducing computational resources.
BE
Bernardo
1 year ago
B: I think using Amazon SageMaker Debugger to stop training jobs when non-converging conditions are detected is also a good idea.
LA
Lashawn
1 year ago
A: I agree, using AWS Trainium instances for training would definitely help reduce energy usage.
ER
Ernest
1 year ago
I'm surprised option C is there. Using Lambda for model deployment? That's like trying to run a marathon with cement shoes!
HO
Honey
1 year ago
Option C is definitely not the best choice for reducing energy usage.
AS
Ashlee
1 year ago
A and E are the way to go! Stopping non-converging jobs and distributed training are key to saving resources.
MI
Michal
1 year ago
I agree, using Amazon SageMaker Debugger can really help in detecting non-converging conditions.
LI
Lizette
1 year ago
A and E are the way to go! Stopping non-converging jobs and distributed training are key to saving resources.
LA
Lasandra
1 year ago
Using PyTorch or TensorFlow with distributed training can be another effective way to reduce computational resources.
AL
Aleta
1 year ago
I believe deploying models with AWS Lambda functions can also help in reducing energy usage.
LO
Lorenza
1 year ago
I agree with Bettina, stopping non-converging conditions early can save resources.
BE
Bettina
1 year ago
I think using Amazon SageMaker Debugger can help reduce energy usage.