transformers.js/pr_1452/en/api/processors.md

# processors

Processors are used to prepare inputs (e.g., text, image or audio) for a model.

**Example:** Using a `WhisperProcessor` to prepare an audio input for a model.
```javascript
import { AutoProcessor, read_audio } from '@huggingface/transformers';

const processor = await AutoProcessor.from_pretrained('openai/whisper-tiny.en');
const audio = await read_audio('https://huggingface.co/datasets/Narsil/asr_dummy/resolve/main/mlk.flac', 16000);
const { input_features } = await processor(audio);
// Tensor {
//   data: Float32Array(240000) [0.4752984642982483, 0.5597258806228638, 0.56434166431427, ...],
//   dims: [1, 80, 3000],
//   type: 'float32',
//   size: 240000,
// }
```


* [processors](#module_processors)
    * _static_
        * [.Processor](#module_processors.Processor)
            * [`new Processor(config, components, chat_template)`](#new_module_processors.Processor_new)
            * _instance_
                * [`.image_processor`](#module_processors.Processor+image_processor) ⇒ <code>*</code>
                * [`.tokenizer`](#module_processors.Processor+tokenizer) ⇒ <code>PreTrainedTokenizer</code> | <code>undefined</code>
                * [`.feature_extractor`](#module_processors.Processor+feature_extractor) ⇒ <code>*</code>
                * [`.apply_chat_template(messages, options)`](#module_processors.Processor+apply_chat_template) ⇒ <code>*</code>
                * [`.batch_decode(...args)`](#module_processors.Processor+batch_decode) ⇒ <code>*</code>
                * [`.decode(...args)`](#module_processors.Processor+decode) ⇒ <code>*</code>
                * [`._call(input, ...args)`](#module_processors.Processor+_call) ⇒ <code>Promise.&lt;any&gt;</code>
            * _static_
                * [`.from_pretrained(pretrained_model_name_or_path, options)`](#module_processors.Processor.from_pretrained) ⇒ <code>Promise.&lt;Processor&gt;</code>
    * _inner_
        * [`~PreTrainedTokenizer`](#module_processors..PreTrainedTokenizer) : <code>Object</code>


* * *

<a id="module_processors.Processor" class="group"></a>

## processors.Processor

Represents a Processor that extracts features from an input.

**Kind**: static class of [<code>processors</code>](#module_processors)  

* [.Processor](#module_processors.Processor)
    * [`new Processor(config, components, chat_template)`](#new_module_processors.Processor_new)
    * _instance_
        * [`.image_processor`](#module_processors.Processor+image_processor) ⇒ <code>*</code>
        * [`.tokenizer`](#module_processors.Processor+tokenizer) ⇒ <code>PreTrainedTokenizer</code> | <code>undefined</code>
        * [`.feature_extractor`](#module_processors.Processor+feature_extractor) ⇒ <code>*</code>
        * [`.apply_chat_template(messages, options)`](#module_processors.Processor+apply_chat_template) ⇒ <code>*</code>
        * [`.batch_decode(...args)`](#module_processors.Processor+batch_decode) ⇒ <code>*</code>
        * [`.decode(...args)`](#module_processors.Processor+decode) ⇒ <code>*</code>
        * [`._call(input, ...args)`](#module_processors.Processor+_call) ⇒ <code>Promise.&lt;any&gt;</code>
    * _static_
        * [`.from_pretrained(pretrained_model_name_or_path, options)`](#module_processors.Processor.from_pretrained) ⇒ <code>Promise.&lt;Processor&gt;</code>


* * *

<a id="new_module_processors.Processor_new" class="group"></a>

### `new Processor(config, components, chat_template)`

Creates a new Processor with the given components

<table>
  <thead>
    <tr>
      <th>Param</th><th>Type</th>
    </tr>
  </thead>
  <tbody>
<tr>
    <td>config</td><td><code>Object</code></td>
    </tr><tr>
    <td>components</td><td><code>Record.&lt;string, Object&gt;</code></td>
    </tr><tr>
    <td>chat_template</td><td><code>string</code></td>
    </tr>  </tbody>
</table>


* * *

<a id="module_processors.Processor+image_processor" class="group"></a>

### `processor.image_processor` ⇒ <code>*</code>

**Kind**: instance property of [<code>Processor</code>](#module_processors.Processor)  
**Returns**: <code>*</code> - The image processor of the processor, if it exists.  

* * *

<a id="module_processors.Processor+tokenizer" class="group"></a>

### `processor.tokenizer` ⇒ <code>PreTrainedTokenizer</code> | <code>undefined</code>

**Kind**: instance property of [<code>Processor</code>](#module_processors.Processor)  
**Returns**: <code>PreTrainedTokenizer</code> | <code>undefined</code> - The tokenizer of the processor, if it exists.  

* * *

<a id="module_processors.Processor+feature_extractor" class="group"></a>

### `processor.feature_extractor` ⇒ <code>*</code>

**Kind**: instance property of [<code>Processor</code>](#module_processors.Processor)  
**Returns**: <code>*</code> - The feature extractor of the processor, if it exists.  

* * *

<a id="module_processors.Processor+apply_chat_template" class="group"></a>

### `processor.apply_chat_template(messages, options)` ⇒ <code>*</code>

**Kind**: instance method of [<code>Processor</code>](#module_processors.Processor)  
<table>
  <thead>
    <tr>
      <th>Param</th><th>Type</th>
    </tr>
  </thead>
  <tbody>
<tr>
    <td>messages</td><td><code>*</code></td>
    </tr><tr>
    <td>options</td><td><code>*</code></td>
    </tr>  </tbody>
</table>


* * *

<a id="module_processors.Processor+batch_decode" class="group"></a>

### `processor.batch_decode(...args)` ⇒ <code>*</code>

**Kind**: instance method of [<code>Processor</code>](#module_processors.Processor)  
<table>
  <thead>
    <tr>
      <th>Param</th><th>Type</th>
    </tr>
  </thead>
  <tbody>
<tr>
    <td>...args</td><td><code>*</code></td>
    </tr>  </tbody>
</table>


* * *

<a id="module_processors.Processor+decode" class="group"></a>

### `processor.decode(...args)` ⇒ <code>*</code>

**Kind**: instance method of [<code>Processor</code>](#module_processors.Processor)  
<table>
  <thead>
    <tr>
      <th>Param</th><th>Type</th>
    </tr>
  </thead>
  <tbody>
<tr>
    <td>...args</td><td><code>*</code></td>
    </tr>  </tbody>
</table>


* * *

<a id="module_processors.Processor+_call" class="group"></a>

### `processor._call(input, ...args)` ⇒ <code>Promise.&lt;any&gt;</code>

Calls the feature_extractor function with the given input.

**Kind**: instance method of [<code>Processor</code>](#module_processors.Processor)  
**Returns**: <code>Promise.&lt;any&gt;</code> - A Promise that resolves with the extracted features.  
<table>
  <thead>
    <tr>
      <th>Param</th><th>Type</th><th>Description</th>
    </tr>
  </thead>
  <tbody>
<tr>
    <td>input</td><td><code>any</code></td><td><p>The input to extract features from.</p>
</td>
    </tr><tr>
    <td>...args</td><td><code>any</code></td><td><p>Additional arguments.</p>
</td>
    </tr>  </tbody>
</table>


* * *

<a id="module_processors.Processor.from_pretrained" class="group"></a>

### `Processor.from_pretrained(pretrained_model_name_or_path, options)` ⇒ <code>Promise.&lt;Processor&gt;</code>

Instantiate one of the processor classes of the library from a pretrained model.

The processor class to instantiate is selected based on the `image_processor_type` (or `feature_extractor_type`; legacy)
property of the config object (either passed as an argument or loaded from `pretrained_model_name_or_path` if possible)

**Kind**: static method of [<code>Processor</code>](#module_processors.Processor)  
**Returns**: <code>Promise.&lt;Processor&gt;</code> - A new instance of the Processor class.  
<table>
  <thead>
    <tr>
      <th>Param</th><th>Type</th><th>Description</th>
    </tr>
  </thead>
  <tbody>
<tr>
    <td>pretrained_model_name_or_path</td><td><code>string</code></td><td><p>The name or path of the pretrained model. Can be either:</p>
<ul>
<li>A string, the <em>model id</em> of a pretrained processor hosted inside a model repo on huggingface.co.
Valid model ids can be located at the root-level, like <code>bert-base-uncased</code>, or namespaced under a
user or organization name, like <code>dbmdz/bert-base-german-cased</code>.</li>
<li>A path to a <em>directory</em> containing processor files, e.g., <code>./my_model_directory/</code>.</li>
</ul>
</td>
    </tr><tr>
    <td>options</td><td><code><a href="#PretrainedProcessorOptions">PretrainedProcessorOptions</a></code></td><td><p>Additional options for loading the processor.</p>
</td>
    </tr>  </tbody>
</table>


* * *

<a id="module_processors..PreTrainedTokenizer" class="group"></a>

## `processors~PreTrainedTokenizer` : <code>Object</code>

Additional processor-specific properties.

**Kind**: inner typedef of [<code>processors</code>](#module_processors)  

* * *



<EditOnGithub source="https://github.com/huggingface/transformers.js/blob/main/docs/source/api/processors.md" />