MENU
A machine learning engineer wants to log and deploy a model as an MLflow pyfunc model. They have custom preprocessing that needs to be completed on feature variables prior to fitting the model or computing predictions using that model. They decide to wrap this preprocessing in a custom model class ModelWithPreprocess, where the preprocessing is performed when calling fit and when calling predict. They then log the fitted model of the ModelWithPreprocess class as a pyfunc model.
Which of the following is a benefit of this approach when loading the logged pyfunc model for downstream deployment?
Suggested Answer: D
Get Premium Databricks Machine Learning Professional Questions
as Interactive Web-Based Practice Test or PDF
Real Exam Format
Detailed Explanations
Instant Access
Get Your Practice Test
41 comments · newest first
TA
Tamie
10 months ago
D seems off; this approach definitely impacts deployment.
MA
Madalyn
11 months ago
Option B is also a solid point, but C seems more relevant for deployment.
MO
Monroe
11 months ago
Wait, does this mean we don't need pipelines anymore? Sounds too good to be true!
KA
Kathrine
11 months ago
Totally agree with C! Saves a lot of hassle.
MA
Martina
11 months ago
I think option C is the best choice! Preprocessing during predict is key.
AN
Annelle
12 months ago
I wonder if option D could be correct, but it seems unlikely that this approach wouldn't have any impact on deployment.
GL
Gladis
12 months ago
I feel like we had a practice question about logging models with custom classes, and it emphasized the need for preprocessing during fit and predict.
KE
Keshia
12 months ago
I'm not entirely sure, but I think option C makes sense since it mentions applying the same preprocessing logic when calling predict.
CA
Candra
1 year ago
I remember we discussed how important it is to ensure that preprocessing is consistent during both training and prediction phases.
LA
Lavina
1 year ago
I'm still a little unsure about the other options. Is there really no benefit to parallelizing the pyfunc model deployment, or is that just not the main advantage here? I want to make sure I fully understand the tradeoffs.
ME
Melodie
1 year ago
Nice, that makes sense. I like how wrapping the preprocessing in the custom model class simplifies the deployment process. It's a clever way to ensure the preprocessing is always applied correctly.
KA
Kasandra
1 year ago
Okay, I think I've got it. The benefit is that the same preprocessing logic will automatically be applied when calling predict on the pyfunc model. That way, I don't have to worry about applying the preprocessing separately when deploying the model.
TO
Tori
1 year ago
I'm a bit confused by the question. Is the benefit that the preprocessing is automatically applied when calling fit and predict on the pyfunc model? Or is there some other benefit I'm missing?
SA
Sanjuana
1 year ago
This seems like a pretty straightforward question. I think the key is understanding how the custom ModelWithPreprocess class works and how that impacts the pyfunc model deployment.
RE
Reita
1 year ago
This question seems straightforward, but I want to make sure I understand the key differences between Vlocity Guided Selling and the regular Vlocity Cart.
GU
Gussie
1 year ago
Okay, let's think this through step-by-step. We need to make sure all devices in area 0 can reach PE1, and the default route needs to be injected. I think Option C might be the way to go, but I'll double-check the other options just to be sure.
EM
Emmett
1 year ago
Wow, this engineer really likes to get their hands dirty, huh? Wrapping the preprocessing in a custom model class? That's some next-level stuff. But hey, if it works, it works. I'm just glad I don't have to worry about all that technical mumbo-jumbo. I'll just stick to the multiple-choice questions and leave the coding to the experts!
CI
Ciara
1 year ago
C) The same preprocessing logic will automatically be applied when calling predict
CE
Cecil
1 year ago
B) The same preprocessing logic will automatically be applied when calling fit
LA
Lasandra
1 year ago
A) The pvfunc model can be used to deploy models in a parallelizable fashion
OD
Odette
1 year ago
Well, this is an interesting approach. I can see the benefit of having the preprocessing baked into the model itself. It'll definitely make deployment a lot simpler. Although, I wonder if it might be a bit of a hassle to maintain that custom model class down the line. Ah, the joys of machine learning engineering!
CE
Celestina
1 year ago
Mabel: True, but it's all part of the fun of machine learning engineering!
MA
Mabel
1 year ago
User 2: It might be a bit tricky to maintain that custom model class though.
LO
Lorita
1 year ago
User 1: I agree, having the preprocessing built into the model does simplify deployment.
WI
Winfred
1 year ago
Francine: True, but I can see how it might become a bit of a headache to maintain that custom model class
FR
Francine
1 year ago
User 2: Definitely, it's a great way to ensure consistency in the preprocessing step
AL
Alline
1 year ago
User 1: I agree, having the preprocessing built into the model does simplify things
AN
Anastacia
1 year ago
Haha, this sounds like the engineer wanted to reinvent the wheel. Why not just use a good old pipeline? That's what they're designed for, to handle all the preprocessing and modeling in one neat package. But I guess everyone has their own style, right?
SY
Sylvia
1 year ago
In the end, as long as the model works effectively, that's what matters.
NA
Nathan
1 year ago
True, everyone has their own approach to solving problems.
SA
Santos
1 year ago
Maybe the engineer wanted more control over the preprocessing steps.
DO
Dorothy
1 year ago
I agree, using a pipeline would have been simpler and more efficient.
MA
Margart
1 year ago
Hmm, I'm not sure if this approach is really necessary. Doesn't MLflow already handle the preprocessing for you when you log a model? Oh well, I guess the engineer wanted to be extra thorough. As long as it works, I'm happy!
MA
Marjory
2 years ago
Ah, I see! So the preprocessing is baked right into the model itself. That's clever, saves me the hassle of managing a separate preprocessing step. Now I can just load up the pyfunc model and let it handle everything. Neat!
EU
Eugene
1 year ago
Antione: And it's great for downstream deployment too. Everything is taken care of within the model itself.
EL
Eladia
1 year ago
User 3: It definitely streamlines the process. Just load the pyfunc model and you're good to go.
AN
Antione
1 year ago
User 2: That's so convenient. No need to worry about separate preprocessing steps anymore.
MA
Malcom
1 year ago
User 1: Yes, exactly! The preprocessing is automatically applied when calling fit and predict.
VI
Virgie
2 years ago
I believe option C is the correct answer then. It makes the deployment process smoother.
TA
Tamekia
2 years ago
I agree with Kanisha. It's convenient to have the preprocessing automatically applied during prediction.
KA
Kanisha
2 years ago
I think the benefit of this approach is that the same preprocessing logic will automatically be applied when calling predict.