Transformers.js documentation

generation/logits_sampler

You are viewing main version, which requires installation from source. If you'd like regular npm install, checkout the latest stable version (v3.0.0).
Hugging Face's logo
Join the Hugging Face community

and get access to the augmented documentation experience

to get started

generation/logits_sampler


generation/logits_sampler.LogitsSampler

Sampler is a base class for all sampling methods used for text generation.

Kind: static class of generation/logits_sampler


new LogitsSampler(generation_config)

Creates a new Sampler object with the specified generation config.

ParamTypeDescription
generation_configGenerationConfig

The generation config.


logitsSampler._call(logits) β‡’ <code> * </code>

Executes the sampler, using the specified logits.

Kind: instance method of LogitsSampler

ParamType
logitsTensor

logitsSampler.sample(logits) β‡’ <code> * </code>

Abstract method for sampling the logits.

Kind: instance method of LogitsSampler
Throws:

  • Error If not implemented in subclass.
ParamType
logitsTensor

logitsSampler.getLogits(logits, index) β‡’ <code> Float32Array </code>

Returns the specified logits as an array, with temperature applied.

Kind: instance method of LogitsSampler

ParamType
logitsTensor
indexnumber

logitsSampler.randomSelect(probabilities) β‡’ <code> number </code>

Selects an item randomly based on the specified probabilities.

Kind: instance method of LogitsSampler
Returns: number - The index of the selected item.

ParamTypeDescription
probabilities*

An array of probabilities to use for selection.


LogitsSampler.getSampler(generation_config) β‡’ <code> LogitsSampler </code>

Returns a Sampler object based on the specified options.

Kind: static method of LogitsSampler
Returns: LogitsSampler - A Sampler object.

ParamTypeDescription
generation_configGenerationConfig

An object containing options for the sampler.


generation/logits_sampler~GreedySampler

Class representing a Greedy Sampler.

Kind: inner class of generation/logits_sampler


greedySampler.sample(logits) β‡’ <code> * </code>

Sample the maximum probability of a given logits tensor.

Kind: instance method of GreedySampler
Returns: * - An array with a single tuple, containing the index of the maximum value and a meaningless score (since this is a greedy search).

ParamType
logitsTensor

generation/logits_sampler~MultinomialSampler

Class representing a MultinomialSampler.

Kind: inner class of generation/logits_sampler


multinomialSampler.sample(logits) β‡’ <code> * </code>

Sample from the logits.

Kind: instance method of MultinomialSampler

ParamType
logitsTensor

generation/logits_sampler~BeamSearchSampler

Class representing a BeamSearchSampler.

Kind: inner class of generation/logits_sampler


beamSearchSampler.sample(logits) β‡’ <code> * </code>

Sample from the logits.

Kind: instance method of BeamSearchSampler

ParamType
logitsTensor

< > Update on GitHub