Hyperopt - Revision history

Admin: /* Example */

2017-10-17T21:18:55Z

Example

← Older revision		Revision as of 21:18, 17 October 2017
Line 45:		Line 45:
	Simple example to optimize a 2D quadratic function: <math>x^2 + y^2</math>		Simple example to optimize a 2D quadratic function: <math>x^2 + y^2</math>

	==Resources		==Resources==

	Scipy talk on Hyperopt: https://youtube.com/watch?v=Mp1xnPfE4PY		Scipy talk on Hyperopt: https://youtube.com/watch?v=Mp1xnPfE4PY

			Github notebook on optimizing a Keras network with Hyperopt: https://github.com/Vooban/Hyperopt-Keras-CNN-CIFAR-100

Admin: /* Example */

2017-10-17T21:18:11Z

Example

← Older revision		Revision as of 21:18, 17 October 2017
Line 45:		Line 45:
	Simple example to optimize a 2D quadratic function: <math>x^2 + y^2</math>		Simple example to optimize a 2D quadratic function: <math>x^2 + y^2</math>

			==Resources

			Scipy talk on Hyperopt: https://youtube.com/watch?v=Mp1xnPfE4PY

Admin: /* Usage */

2017-10-17T21:17:41Z

Usage

← Older revision		Revision as of 21:17, 17 October 2017
Line 33:		Line 33:
	Random search over a space		Random search over a space

	~~Notion of~~ a score - you need some way to score the results		Start by defining a distribution space

			Create a score over the space - you need some way to score the results (your function of interest)

			Then, call fmin to minimize the function over your distribution space

			Specify max_evals to tell it how many random points to try

	===Example===		===Example===

Admin: /* Usage */

2017-10-17T21:16:37Z

Usage

← Older revision		Revision as of 21:16, 17 October 2017
Line 30:		Line 30:

	==Usage==		==Usage==

			Random search over a space

	Notion of a score - you need some way to score the results		Notion of a score - you need some way to score the results

Admin at 21:16, 17 October 2017

2017-10-17T21:16:00Z

@@ Line 2: / Line 2: @@
 Origin of the work was in searching through parameter spaces/sampling probability distributions of parameters.
 Notion of a score - you need some way to score the results
 Simple example to optimize a 2D quadratic function: <math>x^2 + y^2</math>

Admin: Created page with "Hyperopt is a hyperparameter optimization library. Origin of the work was in searching through parameter spaces/sampling probability distributions of parameters. Notion of..."

2017-10-17T21:09:22Z

Created page with "Hyperopt is a hyperparameter optimization library. Origin of the work was in searching through parameter spaces/sampling probability distributions of parameters. Notion of..."

New page

Hyperopt is a hyperparameter optimization library.

Origin of the work was in searching through parameter spaces/sampling probability distributions of parameters.

Notion of a score - you need some way to score the results

Simple example to optimize a 2D quadratic function: <math>x^2 + y^2</math>

[[Category:ML]]
[[Category:NN]]