Combination of all parameters for grid search

Question

I'm training a FastText word-representation model, and I'm performing a grid search over a range of multiple parameters. Below is multiple parameter lists:

wordNgrams = [2, 3, 4, 5]
lr = [10e-2, 10e-3, 10e-4, 10e-5, 10e-6]
dim = [200, 250, 300]
ws = [5, 6, 8, 10]

I'd like to try every possible combination for the above lists together and pass them as parameters to train my model. I don't know how to implement a python function to do that and was hoping for some help. Please help.

score 3 · Accepted Answer · answered Feb 25 '19 at 03:50

3

IIUC use itertools:

from itertools import product
print(product(wordNgrams, lr, dim, ws,repeat=4))

answered Feb 25 '19 at 03:50

U13-Forward

69,221
14
89
114

Thanks for the answer. I just wanted clarification on the repeat parameter, what does it do exactly? – Sam Feb 25 '19 at 17:05
@Sam Your welcome, `repeat` means how many values per subtuple. – U13-Forward Feb 26 '19 at 02:13

score 0 · Answer 2 · answered Jun 18 '21 at 07:00

0

You could also use ParameterGrid from sklearn.model_selection. Explained better in answer to a similar question here.

answered Jun 18 '21 at 07:00

Xichan

95
1
11

Combination of all parameters for grid search

2 Answers2