@min-pack/tfjs-node 2.17.2 → 3.8.1-patch.0

package/tfjs-types/models/segformer/image_processing_segformer.d.ts

@@ -0,0 +1,10 @@
+export class SegformerImageProcessor extends ImageProcessor {
+    post_process_semantic_segmentation(outputs: any, target_sizes?: [number, number][]): {
+        segmentation: import("../../transformers.js").Tensor;
+        labels: number[];
+    }[];
+}
+export class SegformerFeatureExtractor extends SegformerImageProcessor {
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_segformer.d.ts.map

package/tfjs-types/models/siglip/image_processing_siglip.d.ts

@@ -0,0 +1,4 @@
+export class SiglipImageProcessor extends ImageProcessor {
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_siglip.d.ts.map

package/tfjs-types/models/smolvlm/image_processing_smolvlm.d.ts

@@ -0,0 +1,2 @@
+export { Idefics3ImageProcessor as SmolVLMImageProcessor } from "../idefics3/image_processing_idefics3.js";
+//# sourceMappingURL=image_processing_smolvlm.d.ts.map

package/tfjs-types/models/smolvlm/processing_smolvlm.d.ts

@@ -0,0 +1,2 @@
+export { Idefics3Processor as SmolVLMProcessor } from "../idefics3/processing_idefics3.js";
+//# sourceMappingURL=processing_smolvlm.d.ts.map

package/tfjs-types/models/snac/feature_extraction_snac.d.ts

@@ -0,0 +1,4 @@
+export class SnacFeatureExtractor extends DacFeatureExtractor {
+}
+import { DacFeatureExtractor } from '../dac/feature_extraction_dac.js';
+//# sourceMappingURL=feature_extraction_snac.d.ts.map

package/tfjs-types/models/speecht5/feature_extraction_speecht5.d.ts

@@ -0,0 +1,4 @@
+export class SpeechT5FeatureExtractor extends FeatureExtractor {
+}
+import { FeatureExtractor } from "../../base/feature_extraction_utils.js";
+//# sourceMappingURL=feature_extraction_speecht5.d.ts.map

package/tfjs-types/models/speecht5/processing_speecht5.d.ts

@@ -0,0 +1,14 @@
+export class SpeechT5Processor extends Processor {
+    static tokenizer_class: typeof AutoTokenizer;
+    static feature_extractor_class: typeof AutoFeatureExtractor;
+    /**
+     * Calls the feature_extractor function with the given input.
+     * @param {any} input The input to extract features from.
+     * @returns {Promise<any>} A Promise that resolves with the extracted features.
+     */
+    _call(input: any): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+import { AutoFeatureExtractor } from "../auto/feature_extraction_auto.js";
+//# sourceMappingURL=processing_speecht5.d.ts.map

package/tfjs-types/models/swin2sr/image_processing_swin2sr.d.ts

@@ -0,0 +1,5 @@
+export class Swin2SRImageProcessor extends ImageProcessor {
+    pad_image(pixelData: any, imgDims: any, padSize: any, options?: {}): [Float32Array<ArrayBufferLike>, number[]];
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_swin2sr.d.ts.map

package/tfjs-types/models/ultravox/processing_ultravox.d.ts

@@ -0,0 +1,16 @@
+/**
+ * Represents a UltravoxProcessor that extracts features from an audio input.
+ */
+export class UltravoxProcessor extends Processor {
+    static tokenizer_class: typeof AutoTokenizer;
+    static feature_extractor_class: typeof AutoFeatureExtractor;
+    /**
+     * @param {string} text The text input to process.
+     * @param {Float32Array} audio The audio input to process.
+     */
+    _call(text: string, audio?: Float32Array, kwargs?: {}): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+import { AutoFeatureExtractor } from "../auto/feature_extraction_auto.js";
+//# sourceMappingURL=processing_ultravox.d.ts.map

package/tfjs-types/models/vit/image_processing_vit.d.ts

@@ -0,0 +1,6 @@
+export class ViTImageProcessor extends ImageProcessor {
+}
+export class ViTFeatureExtractor extends ViTImageProcessor {
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_vit.d.ts.map

package/tfjs-types/models/vitmatte/image_processing_vitmatte.d.ts

@@ -0,0 +1,12 @@
+export class VitMatteImageProcessor extends ImageProcessor {
+    /**
+     * Calls the feature extraction process on an array of images, preprocesses
+     * each image, and concatenates the resulting features into a single Tensor.
+     * @param {import("../../utils/image.js").RawImage[]} images The image(s) to extract features from.
+     * @param {import("../../utils/image.js").RawImage[]} trimaps The trimaps(s) to extract features from.
+     * @returns {Promise<import("../../base/image_processors_utils.js").ImageProcessorResult>} An object containing the concatenated pixel values of the preprocessed images.
+     */
+    _call(images: import("../../utils/image.js").RawImage[], trimaps: import("../../utils/image.js").RawImage[]): Promise<import("../../base/image_processors_utils.js").ImageProcessorResult>;
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_vitmatte.d.ts.map

package/tfjs-types/models/vitpose/image_processing_vitpose.d.ts

@@ -0,0 +1,26 @@
+export class VitPoseImageProcessor extends ImageProcessor {
+    /**
+     * Transform the heatmaps into keypoint predictions and transform them back to the image.
+     * NOTE: This is a naive implementation and does not include advanced post-processing techniques,
+     * so the results may not be as accurate as the original implementation.
+     * @param {import('../../utils/tensor.js').Tensor} outputs The model outputs.
+     * @param {[number, number, number, number][][]} boxes List or array of bounding boxes for each image.
+     * Each box should be a list of 4 floats representing the bounding box coordinates in COCO format (top_left_x, top_left_y, width, height).
+     * @returns {{
+     *   bbox: [number, number, number, number],
+     *   scores: number[],
+     *   labels: number[],
+     *   keypoints: [number, number][]
+     * }[][]} List of keypoints predictions for each image.
+     */
+    post_process_pose_estimation(outputs: import("../../utils/tensor.js").Tensor, boxes: [number, number, number, number][][], { threshold, }?: {
+        threshold?: any;
+    }): {
+        bbox: [number, number, number, number];
+        scores: number[];
+        labels: number[];
+        keypoints: [number, number][];
+    }[][];
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_vitpose.d.ts.map

package/tfjs-types/models/voxtral/processing_voxtral.d.ts

@@ -0,0 +1,16 @@
+/**
+ * Represents a VoxtralProcessor that extracts features from an audio input.
+ */
+export class VoxtralProcessor extends Processor {
+    static tokenizer_class: typeof AutoTokenizer;
+    static feature_extractor_class: typeof AutoFeatureExtractor;
+    /**
+     * @param {string} text The text input to process.
+     * @param {Float32Array|Float32Array[]} audio The audio input(s) to process.
+     */
+    _call(text: string, audio?: Float32Array | Float32Array[], kwargs?: {}): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+import { AutoFeatureExtractor } from "../auto/feature_extraction_auto.js";
+//# sourceMappingURL=processing_voxtral.d.ts.map

package/tfjs-types/models/wav2vec2/feature_extraction_wav2vec2.d.ts

@@ -0,0 +1,19 @@
+export class Wav2Vec2FeatureExtractor extends FeatureExtractor {
+    /**
+     * @param {Float32Array} input_values
+     * @returns {Float32Array}
+     */
+    _zero_mean_unit_var_norm(input_values: Float32Array): Float32Array;
+    /**
+     * Asynchronously extracts features from a given audio using the provided configuration.
+     * @param {Float32Array|Float64Array} audio The audio data as a Float32Array/Float64Array.
+     * @returns {Promise<{ input_values: Tensor; attention_mask: Tensor }>} A Promise resolving to an object containing the extracted input features and attention mask as Tensors.
+     */
+    _call(audio: Float32Array | Float64Array): Promise<{
+        input_values: Tensor;
+        attention_mask: Tensor;
+    }>;
+}
+import { FeatureExtractor } from "../../base/feature_extraction_utils.js";
+import { Tensor } from "../../utils/tensor.js";
+//# sourceMappingURL=feature_extraction_wav2vec2.d.ts.map

package/tfjs-types/models/wav2vec2/processing_wav2vec2.d.ts

@@ -0,0 +1,14 @@
+export class Wav2Vec2Processor extends Processor {
+    static tokenizer_class: typeof AutoTokenizer;
+    static feature_extractor_class: typeof AutoFeatureExtractor;
+    /**
+     * Calls the feature_extractor function with the given audio input.
+     * @param {any} audio The audio input to extract features from.
+     * @returns {Promise<any>} A Promise that resolves with the extracted features.
+     */
+    _call(audio: any): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+import { AutoFeatureExtractor } from "../auto/feature_extraction_auto.js";
+//# sourceMappingURL=processing_wav2vec2.d.ts.map

package/tfjs-types/models/wav2vec2_with_lm/processing_wav2vec2_with_lm.d.ts

@@ -0,0 +1,14 @@
+export class Wav2Vec2ProcessorWithLM extends Processor {
+    static tokenizer_class: typeof AutoTokenizer;
+    static feature_extractor_class: typeof AutoFeatureExtractor;
+    /**
+     * Calls the feature_extractor function with the given audio input.
+     * @param {any} audio The audio input to extract features from.
+     * @returns {Promise<any>} A Promise that resolves with the extracted features.
+     */
+    _call(audio: any): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+import { AutoFeatureExtractor } from "../auto/feature_extraction_auto.js";
+//# sourceMappingURL=processing_wav2vec2_with_lm.d.ts.map

package/tfjs-types/models/wespeaker/feature_extraction_wespeaker.d.ts

@@ -0,0 +1,23 @@
+export class WeSpeakerFeatureExtractor extends FeatureExtractor {
+    constructor(config: any);
+    mel_filters: number[][];
+    window: Float64Array<ArrayBufferLike>;
+    min_num_frames: any;
+    /**
+     * Computes the log-Mel spectrogram of the provided audio waveform.
+     * @param {Float32Array|Float64Array} waveform The audio waveform to process.
+     * @returns {Promise<Tensor>} An object containing the log-Mel spectrogram data as a Float32Array and its dimensions as an array of numbers.
+     */
+    _extract_fbank_features(waveform: Float32Array | Float64Array): Promise<Tensor>;
+    /**
+     * Asynchronously extracts features from a given audio using the provided configuration.
+     * @param {Float32Array|Float64Array} audio The audio data as a Float32Array/Float64Array.
+     * @returns {Promise<{ input_features: Tensor }>} A Promise resolving to an object containing the extracted input features as a Tensor.
+     */
+    _call(audio: Float32Array | Float64Array): Promise<{
+        input_features: Tensor;
+    }>;
+}
+import { FeatureExtractor } from '../../base/feature_extraction_utils.js';
+import { Tensor } from '../../utils/tensor.js';
+//# sourceMappingURL=feature_extraction_wespeaker.d.ts.map

package/tfjs-types/models/whisper/common_whisper.d.ts

@@ -0,0 +1,8 @@
+/**
+ * @param {string} language The language name or code
+ * @returns {string} The language code
+ */
+export function whisper_language_to_code(language: string): string;
+export const WHISPER_LANGUAGE_MAPPING: Map<any, any>;
+export const WHISPER_TO_LANGUAGE_CODE_MAPPING: Map<any, any>;
+//# sourceMappingURL=common_whisper.d.ts.map

package/tfjs-types/models/whisper/feature_extraction_whisper.d.ts

@@ -0,0 +1,23 @@
+export class WhisperFeatureExtractor extends FeatureExtractor {
+    constructor(config: any);
+    window: Float64Array<ArrayBufferLike>;
+    /**
+     * Computes the log-Mel spectrogram of the provided audio waveform.
+     * @param {Float32Array|Float64Array} waveform The audio waveform to process.
+     * @returns {Promise<Tensor>} An object containing the log-Mel spectrogram data as a Float32Array and its dimensions as an array of numbers.
+     */
+    _extract_fbank_features(waveform: Float32Array | Float64Array): Promise<Tensor>;
+    /**
+     * Asynchronously extracts features from a given audio using the provided configuration.
+     * @param {Float32Array|Float64Array} audio The audio data as a Float32Array/Float64Array.
+     * @returns {Promise<{ input_features: Tensor }>} A Promise resolving to an object containing the extracted input features as a Tensor.
+     */
+    _call(audio: Float32Array | Float64Array, { max_length, }?: {
+        max_length?: any;
+    }): Promise<{
+        input_features: Tensor;
+    }>;
+}
+import { FeatureExtractor } from '../../base/feature_extraction_utils.js';
+import { Tensor } from '../../utils/tensor.js';
+//# sourceMappingURL=feature_extraction_whisper.d.ts.map

package/tfjs-types/models/whisper/generation_whisper.d.ts

@@ -0,0 +1,76 @@
+export class WhisperGenerationConfig extends GenerationConfig {
+    /**
+     * Whether to return the timestamps with the text. This enables the `WhisperTimestampsLogitsProcessor`.
+     * @type {boolean}
+     */
+    return_timestamps: boolean;
+    /**
+     * Whether to return token-level timestamps
+     * with the text. This can be used with or without the `return_timestamps` option. To get word-level
+     * timestamps, use the tokenizer to group the tokens into words.
+     * @type {boolean}
+     */
+    return_token_timestamps: boolean;
+    /**
+     * The number of audio frames available in this chunk. This is only used generating word-level timestamps.
+     * @type {number}
+     */
+    num_frames: number;
+    /**
+     * Alignment heads to predict word-level timestamps. This is a list of [layer, head] pairs that
+     * select the cross-attention heads that are highly correlated to word-level timing.
+     * @type {[number, number][]}
+     */
+    alignment_heads: [number, number][];
+    /**
+     * Task to use for generation, either "translate" or "transcribe".
+     * @type {string}
+     */
+    task: string;
+    /**
+     * Language token to use for generation, can be either in the form of `<|en|>`, `en` or `english`.
+     * You can find all the possible language tokens in the `model.generation_config.lang_to_id` dictionary.
+     * @type {string}
+     */
+    language: string;
+    /**
+     * The id of the `"<|notimestamps|>"` token.
+     * @type {number}
+     */
+    no_timestamps_token_id: number;
+    /**
+     * Rank-1 list of token IDs created by passing text to [`~WhisperProcessor.get_prompt_ids`] that is
+     * provided as a prompt to each chunk. This can be used to provide or "prompt-engineer" a context for
+     * transcription, e.g. custom vocabularies or proper nouns to make it more likely to predict those words
+     * correctly. It cannot be used in conjunction with `decoder_start_token_id` as it overwrites this value.
+     * @type {number[]}
+     */
+    prompt_ids: number[];
+    /**
+     * Whether the model is multilingual or not.
+     * @type {boolean}
+     */
+    is_multilingual: boolean;
+    /**
+     * (Optional) A mapping from language tokens to their corresponding IDs.
+     * Only required if the model is multilingual.
+     * @type {Record<string, number>|null}
+     */
+    lang_to_id: Record<string, number> | null;
+    /**
+     * (Optional) A mapping from task tokens to their corresponding IDs.
+     * @type {Record<string, number>|null}
+     */
+    task_to_id: Record<string, number> | null;
+    /**
+     * Used to set the maximum value of the initial timestamp. This is used to prevent the model from
+     * predicting timestamps that are too far in the future.
+     * @type {number}
+     */
+    max_initial_timestamp_index: number;
+}
+export type WhisperGenerationFunctionParameters = import("../../generation/parameters.js").GenerationFunctionParameters & {
+    generation_config: WhisperGenerationConfig;
+} & WhisperGenerationConfig;
+import { GenerationConfig } from "../../generation/configuration_utils.js";
+//# sourceMappingURL=generation_whisper.d.ts.map

package/tfjs-types/models/whisper/processing_whisper.d.ts

@@ -0,0 +1,17 @@
+/**
+ * Represents a WhisperProcessor that extracts features from an audio input.
+ */
+export class WhisperProcessor extends Processor {
+    static tokenizer_class: typeof AutoTokenizer;
+    static feature_extractor_class: typeof AutoFeatureExtractor;
+    /**
+     * Calls the feature_extractor function with the given audio input.
+     * @param {any} audio The audio input to extract features from.
+     * @returns {Promise<any>} A Promise that resolves with the extracted features.
+     */
+    _call(audio: any): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+import { AutoFeatureExtractor } from "../auto/feature_extraction_auto.js";
+//# sourceMappingURL=processing_whisper.d.ts.map

package/tfjs-types/models/yolos/image_processing_yolos.d.ts

@@ -0,0 +1,10 @@
+export class YolosImageProcessor extends ImageProcessor {
+    post_process_object_detection(outputs: {
+        logits: import("../../transformers.js").Tensor;
+        pred_boxes: import("../../transformers.js").Tensor;
+    }, threshold?: number, target_sizes?: [number, number][], is_zero_shot?: boolean): any[];
+}
+export class YolosFeatureExtractor extends YolosImageProcessor {
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_yolos.d.ts.map