Skip to content

Latest commit

 

History

History
101 lines (70 loc) · 7.57 KB

SearchSpaceSpec.md

File metadata and controls

101 lines (70 loc) · 7.57 KB

Search Space

Overview

In NNI, tuner will sample parameters/architecture according to the search space, which is defined as a json file.

To define a search space, users should define the name of variable, the type of sampling strategy and its parameters.

  • An example of search space definition as follow:
{
    "dropout_rate": {"_type": "uniform", "_value": [0.1, 0.5]},
    "conv_size": {"_type": "choice", "_value": [2, 3, 5, 7]},
    "hidden_size": {"_type": "choice", "_value": [124, 512, 1024]},
    "batch_size": {"_type": "choice", "_value": [50, 250, 500]},
    "learning_rate": {"_type": "uniform", "_value": [0.0001, 0.1]}
}

Take the first line as an example. dropout_rate is defined as a variable whose priori distribution is a uniform distribution of a range from 0.1 and 0.5.

Note that the ability of a search space is highly connected with your tuner. We listed the supported types for each builtin tuner below. For a customized tuner, you don't have to follow our convention and you will have the flexibility to define any type you want.

Types

All types of sampling strategies and their parameter are listed here:

  • {"_type": "choice", "_value": options}

    • Which means the variable's value is one of the options. Here options should be a list of numbers or a list of strings. Using arbitrary objects as members of this list (like sublists, a mixture of numbers and strings, or null values) should work in most cases, but may trigger undefined behaviors.

    • options could also be a nested sub-search-space, this sub-search-space takes effect only when the corresponding element is chosen. The variables in this sub-search-space could be seen as conditional variables. Here is an simple example of nested search space definition. If an element in the options list is a dict, it is a sub-search-space, and for our built-in tuners you have to add a key _name in this dict, which helps you to identify which element is chosen. Accordingly, here is a sample which users can get from nni with nested search space definition. Tuners which support nested search space are as follows:

      • Random Search
      • TPE
      • Anneal
      • Evolution
  • {"_type": "randint", "_value": [lower, upper]}

    • Choosing a random integer from lower (inclusive) to upper (exclusive).
    • Note: Different tuners may interpret randint differently. Some (e.g., TPE, GridSearch) treat integers from lower to upper as unordered ones, while others respect the ordering (e.g., SMAC). If you want all the tuners to respect the ordering, please use quniform with q=1.
  • {"_type": "uniform", "_value": [low, high]}

    • Which means the variable value is a value uniformly between low and high.
    • When optimizing, this variable is constrained to a two-sided interval.
  • {"_type": "quniform", "_value": [low, high, q]}

    • Which means the variable value is a value like clip(round(uniform(low, high) / q) * q, low, high), where the clip operation is used to constraint the generated value in the bound. For example, for _value specified as [0, 10, 2.5], possible values are [0, 2.5, 5.0, 7.5, 10.0]; For _value specified as [2, 10, 5], possible values are [2, 5, 10].
    • Suitable for a discrete value with respect to which the objective is still somewhat "smooth", but which should be bounded both above and below. If you want to uniformly choose integer from a range [low, high], you can write _value like this: [low, high, 1].
  • {"_type": "loguniform", "_value": [low, high]}

    • Which means the variable value is a value drawn from a range [low, high] according to a loguniform distribution like exp(uniform(log(low), log(high))), so that the logarithm of the return value is uniformly distributed.
    • When optimizing, this variable is constrained to be positive.
  • {"_type": "qloguniform", "_value": [low, high, q]}

    • Which means the variable value is a value like clip(round(loguniform(low, high) / q) * q, low, high), where the clip operation is used to constraint the generated value in the bound.
    • Suitable for a discrete variable with respect to which the objective is "smooth" and gets smoother with the size of the value, but which should be bounded both above and below.
  • {"_type": "normal", "_value": [mu, sigma]}

    • Which means the variable value is a real value that's normally-distributed with mean mu and standard deviation sigma. When optimizing, this is an unconstrained variable.
  • {"_type": "qnormal", "_value": [mu, sigma, q]}

    • Which means the variable value is a value like round(normal(mu, sigma) / q) * q
    • Suitable for a discrete variable that probably takes a value around mu, but is fundamentally unbounded.
  • {"_type": "lognormal", "_value": [mu, sigma]}

    • Which means the variable value is a value drawn according to exp(normal(mu, sigma)) so that the logarithm of the return value is normally distributed. When optimizing, this variable is constrained to be positive.
  • {"_type": "qlognormal", "_value": [mu, sigma, q]}

    • Which means the variable value is a value like round(exp(normal(mu, sigma)) / q) * q
    • Suitable for a discrete variable with respect to which the objective is smooth and gets smoother with the size of the variable, which is bounded from one side.

Search Space Types Supported by Each Tuner

choice randint uniform quniform loguniform qloguniform normal qnormal lognormal qlognormal
TPE Tuner
Random Search Tuner
Anneal Tuner
Evolution Tuner
SMAC Tuner
Batch Tuner
Grid Search Tuner
Hyperband Advisor
Metis Tuner
GP Tuner

Known Limitations:

  • GP Tuner and Metis Tuner support only numerical values in search space (choice type values can be no-numeraical with other tuners, e.g. string values). Both GP Tuner and Metis Tuner use Gaussian Process Regressor(GPR). GPR make predictions based on a kernel function and the 'distance' between different points, it's hard to get the true distance between no-numerical values.

  • Note that for nested search space:

    • Only Random Search/TPE/Anneal/Evolution tuner supports nested search space

    • We do not support nested search space "Hyper Parameter" in visualization now, the enhancement is being considered in #1110, any suggestions or discussions or contributions are warmly welcomed