Nested parallelism with scikit learn models

Question

I want to to do nested parallelism with scikit learn logisticregressionCV inside a for loop:

for i in range(0,10):
    logisticregressionCV(n_jobs=-1)

I want to parallelize the for loop as well.

I read a lot of post but I couldn't understand much. one of the post is this and this.

NOTE: This 'for' loop is not for cross-validation. 'i' can be thought of another dataset obtained from another set of commands not mentioned here i.e i have 10 different datasets and i compute logistic regression for each dataset parallelly.

Please help! have been stuck on this problem for the past 2 days!

MaxU - stand with Ukraine · Accepted Answer · 2018-07-14 11:30:39Z

1

First of all it's not clear why do you need this loop at all. Instead of having this loop just pass the cv=10 parameter.
Using parameter n_jobs=-1 already takes care of running this job in parallel on all available CPU cores.

answered Jul 14, 2018 at 11:30

MaxU - stand with Ukraine

212k37 gold badges402 silver badges436 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Aakash saboo Over a year ago

hi, this 'for' loop is not for cross-validation, you can think 'i' as another set of data obtained from another set of commands(not mentioned here),i.e there are 10 set of data and I pick one dataset and use logistic regression on it(which itself is parallelised ). I want to parallelise this for loop so that logistic regression is computed for all 10 datasets simultaneously and NOT SEQUENTIALLY.

MaxU - stand with Ukraine Over a year ago

@Aakashsaboo, can you elaborate? What are you trying to do? Why and how are are you going to use LogisticRegressionCV for different pieces of data? What would it bring you? And how are you going to merge results?

Collectives™ on Stack Overflow

Nested parallelism with scikit learn models

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related