@min-pack/tfjs-node 2.17.1 → 3.8.1-patch.0

package/tfjs-types/models/auto/processing_auto.d.ts

@@ -0,0 +1,39 @@
+/**
+ * @typedef {import('../../base/processing_utils.js').PretrainedProcessorOptions} PretrainedProcessorOptions
+ */
+/**
+ * Helper class which is used to instantiate pretrained processors with the `from_pretrained` function.
+ * The chosen processor class is determined by the type specified in the processor config.
+ *
+ * **Example:** Load a processor using `from_pretrained`.
+ * ```javascript
+ * let processor = await AutoProcessor.from_pretrained('openai/whisper-tiny.en');
+ * ```
+ *
+ * **Example:** Run an image through a processor.
+ * ```javascript
+ * let processor = await AutoProcessor.from_pretrained('Xenova/clip-vit-base-patch16');
+ * let image = await RawImage.read('https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/football-match.jpg');
+ * let image_inputs = await processor(image);
+ * // {
+ * //   "pixel_values": {
+ * //     "dims": [ 1, 3, 224, 224 ],
+ * //     "type": "float32",
+ * //     "data": Float32Array [ -1.558687686920166, -1.558687686920166, -1.5440893173217773, ... ],
+ * //     "size": 150528
+ * //   },
+ * //   "original_sizes": [
+ * //     [ 533, 800 ]
+ * //   ],
+ * //   "reshaped_input_sizes": [
+ * //     [ 224, 224 ]
+ * //   ]
+ * // }
+ * ```
+ */
+export class AutoProcessor {
+    static from_pretrained(pretrained_model_name_or_path: string, options?: import("../../base/processing_utils.js").PretrainedProcessorOptions): Promise<Processor>;
+}
+export type PretrainedProcessorOptions = import("../../base/processing_utils.js").PretrainedProcessorOptions;
+import { Processor } from '../../base/processing_utils.js';
+//# sourceMappingURL=processing_auto.d.ts.map

package/tfjs-types/models/beit/image_processing_beit.d.ts

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

package/tfjs-types/models/bit/image_processing_bit.d.ts

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

package/tfjs-types/models/chinese_clip/image_processing_chinese_clip.d.ts

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

package/tfjs-types/models/clap/feature_extraction_clap.d.ts

@@ -0,0 +1,57 @@
+export class ClapFeatureExtractor extends FeatureExtractor {
+    constructor(config: any);
+    mel_filters: number[][];
+    mel_filters_slaney: number[][];
+    window: Float64Array<ArrayBufferLike>;
+    /**
+     * Extracts the mel spectrogram and prepares it for the mode based on the `truncation` and `padding` arguments.
+     *
+     * Four different path are possible:
+     *   - `truncation="fusion"` and the length of the waveform is greater than the max length: the mel spectrogram
+     *     will be computed on the entire audio. 3 random crops and a dowsampled version of the full mel spectrogram
+     *     are then stacked together. They will later be used for `feature_fusion`.
+     *   - `truncation="rand_trunc"` and the length of the waveform is smaller than the max length: the audio is
+     *     padded based on `padding`.
+     *   - `truncation="fusion"` and the length of the waveform is smaller than the max length: the audio is padded
+     *     based on `padding`, and is repeated `4` times.
+     *   - `truncation="rand_trunc"` and the length of the waveform is greater than the max length: the mel
+     *     spectrogram will be computed on a random crop of the waveform.
+     *
+     * @param {Float32Array|Float64Array} waveform The input waveform.
+     * @param {number} max_length The maximum length of the waveform.
+     * @param {string} truncation The truncation strategy to use.
+     * @param {string} padding The padding strategy to use.
+     * @returns {Promise<Tensor>} An object containing the mel spectrogram data as a Float32Array, its dimensions as an array of numbers, and a boolean indicating whether the waveform was longer than the max length.
+     * @private
+     */
+    private _get_input_mel;
+    /**
+     * Compute the log-mel spectrogram of the provided `waveform` using the Hann window.
+     * In CLAP, two different filter banks are used depending on the truncation pattern:
+     *  - `self.mel_filters`: they correspond to the default parameters of `torchaudio` which can be obtained from
+     *    calling `torchaudio.transforms.MelSpectrogram().mel_scale.fb`. These filters are used when `truncation`
+     *    is set to `"fusion"`.
+     *  - `self.mel_filteres_slaney` : they correspond to the default parameters of `librosa` which used
+     *    `librosa.filters.mel` when computing the mel spectrogram. These filters were only used in the original
+     *    implementation when the truncation mode is not `"fusion"`.
+     *
+     * @param {Float32Array|Float64Array} waveform The audio waveform to process.
+     * @param {number[][]} mel_filters The mel filters to use.
+     * @param {number} [max_length=null] The maximum number of frames to return.
+     * @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, mel_filters: number[][], max_length?: number): 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_clap.d.ts.map

package/tfjs-types/models/clip/image_processing_clip.d.ts

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

package/tfjs-types/models/convnext/image_processing_convnext.d.ts

@@ -0,0 +1,12 @@
+export class ConvNextImageProcessor extends ImageProcessor {
+    constructor(config: any);
+    /**
+     * Percentage of the image to crop. Only has an effect if this.size < 384.
+     */
+    crop_pct: any;
+    resize(image: any): Promise<any>;
+}
+export class ConvNextFeatureExtractor extends ConvNextImageProcessor {
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_convnext.d.ts.map

package/tfjs-types/models/dac/feature_extraction_dac.d.ts

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

package/tfjs-types/models/deit/image_processing_deit.d.ts

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

package/tfjs-types/models/detr/image_processing_detr.d.ts

@@ -0,0 +1,42 @@
+/**
+ * @typedef {object} DetrFeatureExtractorResultProps
+ * @property {import('../../utils/tensor.js').Tensor} pixel_mask
+ * @typedef {import('../../base/image_processors_utils.js').ImageProcessorResult & DetrFeatureExtractorResultProps} DetrFeatureExtractorResult
+ */
+export class DetrImageProcessor 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.
+     * @returns {Promise<DetrFeatureExtractorResult>} An object containing the concatenated pixel values of the preprocessed images.
+     */
+    _call(images: import("../../utils/image.js").RawImage[]): Promise<DetrFeatureExtractorResult>;
+    post_process_object_detection(outputs: {
+        logits: import("../../utils/tensor.js").Tensor;
+        pred_boxes: import("../../utils/tensor.js").Tensor;
+    }, threshold?: number, target_sizes?: [number, number][], is_zero_shot?: boolean): any[];
+    post_process_panoptic_segmentation(outputs: any, threshold?: number, mask_threshold?: number, overlap_mask_area_threshold?: number, label_ids_to_fuse?: Set<number>, target_sizes?: [number, number][]): Array<{
+        segmentation: import("../../utils/tensor.js").Tensor;
+        segments_info: Array<{
+            id: number;
+            label_id: number;
+            score: number;
+        }>;
+    }>;
+    post_process_instance_segmentation(outputs: any, threshold?: number, target_sizes?: [number, number][]): Array<{
+        segmentation: import("../../utils/tensor.js").Tensor;
+        segments_info: Array<{
+            id: number;
+            label_id: number;
+            score: number;
+        }>;
+    }>;
+}
+export class DetrFeatureExtractor extends DetrImageProcessor {
+}
+export type DetrFeatureExtractorResultProps = {
+    pixel_mask: import("../../utils/tensor.js").Tensor;
+};
+export type DetrFeatureExtractorResult = import("../../base/image_processors_utils.js").ImageProcessorResult & DetrFeatureExtractorResultProps;
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_detr.d.ts.map

package/tfjs-types/models/dinov3_vit/image_processing_dinov3_vit.d.ts

@@ -0,0 +1,4 @@
+export class DINOv3ViTImageProcessor extends ImageProcessor {
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_dinov3_vit.d.ts.map

package/tfjs-types/models/donut/image_processing_donut.d.ts

@@ -0,0 +1,7 @@
+export class DonutImageProcessor extends ImageProcessor {
+    pad_image(pixelData: any, imgDims: any, padSize: any, options?: {}): [Float32Array<ArrayBufferLike>, number[]];
+}
+export class DonutFeatureExtractor extends DonutImageProcessor {
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_donut.d.ts.map

package/tfjs-types/models/dpt/image_processing_dpt.d.ts

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

package/tfjs-types/models/efficientnet/image_processing_efficientnet.d.ts

@@ -0,0 +1,6 @@
+export class EfficientNetImageProcessor extends ImageProcessor {
+    constructor(config: any);
+    include_top: any;
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_efficientnet.d.ts.map

package/tfjs-types/models/encodec/feature_extraction_encodec.d.ts

@@ -0,0 +1,13 @@
+export class EncodecFeatureExtractor extends FeatureExtractor {
+    /**
+     * Asynchronously extracts input values from a given audio using the provided configuration.
+     * @param {Float32Array|Float64Array} audio The audio data as a Float32Array/Float64Array.
+     * @returns {Promise<{ input_values: Tensor; }>} The extracted input values.
+     */
+    _call(audio: Float32Array | Float64Array): Promise<{
+        input_values: Tensor;
+    }>;
+}
+import { FeatureExtractor } from '../../base/feature_extraction_utils.js';
+import { Tensor } from '../../utils/tensor.js';
+//# sourceMappingURL=feature_extraction_encodec.d.ts.map

package/tfjs-types/models/feature_extractors.d.ts

@@ -0,0 +1,16 @@
+export * from "./audio_spectrogram_transformer/feature_extraction_audio_spectrogram_transformer.js";
+export * from "./encodec/feature_extraction_encodec.js";
+export * from "./clap/feature_extraction_clap.js";
+export * from "./dac/feature_extraction_dac.js";
+export * from "./gemma3n/feature_extraction_gemma3n.js";
+export * from "./moonshine/feature_extraction_moonshine.js";
+export * from "./parakeet/feature_extraction_parakeet.js";
+export * from "./pyannote/feature_extraction_pyannote.js";
+export * from "./seamless_m4t/feature_extraction_seamless_m4t.js";
+export * from "./snac/feature_extraction_snac.js";
+export * from "./speecht5/feature_extraction_speecht5.js";
+export * from "./wav2vec2/feature_extraction_wav2vec2.js";
+export * from "./wespeaker/feature_extraction_wespeaker.js";
+export * from "./whisper/feature_extraction_whisper.js";
+export { ImageProcessor as ImageFeatureExtractor } from "../base/image_processors_utils.js";
+//# sourceMappingURL=feature_extractors.d.ts.map

package/tfjs-types/models/florence2/processing_florence2.d.ts

@@ -0,0 +1,39 @@
+export class Florence2Processor extends Processor {
+    static tokenizer_class: typeof AutoTokenizer;
+    static image_processor_class: typeof AutoImageProcessor;
+    constructor(config: any, components: any, chat_template: any);
+    /** @type {Map<string, string>} */
+    tasks_answer_post_processing_type: Map<string, string>;
+    /** @type {Map<string, string>} */
+    task_prompts_without_inputs: Map<string, string>;
+    /** @type {Map<string, string>} */
+    task_prompts_with_input: Map<string, string>;
+    regexes: {
+        quad_boxes: RegExp;
+        bboxes: RegExp;
+    };
+    size_per_bin: number;
+    /**
+     * Helper function to construct prompts from input texts
+     * @param {string|string[]} text
+     * @returns {string[]}
+     */
+    construct_prompts(text: string | string[]): string[];
+    /**
+     * Post-process the output of the model to each of the task outputs.
+     * @param {string} text The text to post-process.
+     * @param {string} task The task to post-process the text for.
+     * @param {[number, number]} image_size The size of the image. height x width.
+     */
+    post_process_generation(text: string, task: string, image_size: [number, number]): {
+        [task]: string | {
+            [x: string]: any[];
+            labels: any[];
+        };
+    };
+    _call(images: any, text?: any, kwargs?: {}): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+import { AutoImageProcessor } from "../auto/image_processing_auto.js";
+//# sourceMappingURL=processing_florence2.d.ts.map

package/tfjs-types/models/gemma3n/feature_extraction_gemma3n.d.ts

@@ -0,0 +1,35 @@
+export class Gemma3nAudioFeatureExtractor extends FeatureExtractor {
+    constructor(config: any);
+    mel_filters: number[][];
+    window: Float64Array<ArrayBufferLike>;
+    /**
+     * Computes the log-Mel spectrogram of the provided audio waveform.
+     * @param {Float32Array|Float64Array} waveform The audio waveform to process.
+     * @param {number} max_length The maximum number of frames to return.
+     * @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, max_length: number): Promise<Tensor>;
+    /**
+     * Asynchronously extracts features from a given audio using the provided configuration.
+     * @param {Float32Array|Float64Array} audio The audio data as a Float32Array/Float64Array.
+     * @param {Object} options Optional parameters for feature extraction.
+     * @param {number} [options.max_length=480_000] If provided, defines the maximum length of the audio to allow.
+     * Audio longer than this will be truncated if `truncation=True`.
+     * @param {boolean} [options.truncation=true] Whether or not to truncate audio above `max_length`.
+     * @param {boolean} [options.padding=true] Whether to pad the sequence to a multiple of `pad_to_multiple_of`.
+     * @param {number} [options.pad_to_multiple_of=128] The number to pad the sequence to a multiple of.
+     * @returns {Promise<{ input_features: Tensor, input_features_mask: Tensor }>} A Promise resolving to an object containing the extracted input features and attention masks as Tensors.
+     */
+    _call(audio: Float32Array | Float64Array, { max_length, truncation, padding, pad_to_multiple_of, }?: {
+        max_length?: number;
+        truncation?: boolean;
+        padding?: boolean;
+        pad_to_multiple_of?: number;
+    }): Promise<{
+        input_features: Tensor;
+        input_features_mask: Tensor;
+    }>;
+}
+import { FeatureExtractor } from '../../base/feature_extraction_utils.js';
+import { Tensor } from '../../utils/tensor.js';
+//# sourceMappingURL=feature_extraction_gemma3n.d.ts.map

package/tfjs-types/models/gemma3n/processing_gemma3n.d.ts

@@ -0,0 +1,31 @@
+export class Gemma3nProcessor extends Processor {
+    static image_processor_class: typeof AutoImageProcessor;
+    static feature_extractor_class: typeof AutoFeatureExtractor;
+    static tokenizer_class: typeof AutoTokenizer;
+    constructor(config: any, components: any, chat_template: any);
+    audio_seq_length: any;
+    image_seq_length: any;
+    audio_token_id: any;
+    boa_token: any;
+    audio_token: any;
+    full_audio_sequence: string;
+    image_token_id: any;
+    boi_token: any;
+    image_token: any;
+    full_image_sequence: string;
+    /**
+     *
+     * @param {string|string[]} text
+     * @param {RawImage|RawImage[]|RawImage[][]} images
+     * @param {RawAudio|RawAudio[]|RawAudio[][]} audio
+     * @returns {Promise<any>}
+     */
+    _call(text: string | string[], images?: RawImage | RawImage[] | RawImage[][], audio?: RawAudio | RawAudio[] | RawAudio[][], options?: {}): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { RawImage } from "../../utils/image.js";
+import { RawAudio } from "../../utils/audio.js";
+import { AutoImageProcessor } from "../auto/image_processing_auto.js";
+import { AutoFeatureExtractor } from "../auto/feature_extraction_auto.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+//# sourceMappingURL=processing_gemma3n.d.ts.map

package/tfjs-types/models/glpn/image_processing_glpn.d.ts

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

package/tfjs-types/models/grounding_dino/image_processing_grounding_dino.d.ts

@@ -0,0 +1,20 @@
+/**
+ * @typedef {object} GroundingDinoFeatureExtractorResultProps
+ * @property {import('../../utils/tensor.js').Tensor} pixel_mask
+ * @typedef {import('../../base/image_processors_utils.js').ImageProcessorResult & GroundingDinoFeatureExtractorResultProps} GroundingDinoFeatureExtractorResult
+ */
+export class GroundingDinoImageProcessor 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.
+     * @returns {Promise<GroundingDinoFeatureExtractorResult>} An object containing the concatenated pixel values of the preprocessed images.
+     */
+    _call(images: import("../../utils/image.js").RawImage[]): Promise<GroundingDinoFeatureExtractorResult>;
+}
+export type GroundingDinoFeatureExtractorResultProps = {
+    pixel_mask: import("../../utils/tensor.js").Tensor;
+};
+export type GroundingDinoFeatureExtractorResult = import("../../base/image_processors_utils.js").ImageProcessorResult & GroundingDinoFeatureExtractorResultProps;
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_grounding_dino.d.ts.map

package/tfjs-types/models/grounding_dino/processing_grounding_dino.d.ts

@@ -0,0 +1,27 @@
+export class GroundingDinoProcessor extends Processor {
+    static tokenizer_class: typeof AutoTokenizer;
+    static image_processor_class: typeof AutoImageProcessor;
+    /**
+     * @typedef {import('../../utils/image.js').RawImage} RawImage
+     */
+    /**
+     *
+     * @param {RawImage|RawImage[]|RawImage[][]} images
+     * @param {string|string[]} text
+     * @returns {Promise<any>}
+     */
+    _call(images: import("../../utils/image.js").RawImage | import("../../utils/image.js").RawImage[] | import("../../utils/image.js").RawImage[][], text: string | string[], options?: {}): Promise<any>;
+    post_process_grounded_object_detection(outputs: any, input_ids: any, { box_threshold, text_threshold, target_sizes }?: {
+        box_threshold?: number;
+        text_threshold?: number;
+        target_sizes?: any;
+    }): {
+        scores: any[];
+        boxes: any[];
+        labels: string[];
+    }[];
+}
+import { Processor } from "../../base/processing_utils.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+import { AutoImageProcessor } from "../auto/image_processing_auto.js";
+//# sourceMappingURL=processing_grounding_dino.d.ts.map

package/tfjs-types/models/idefics3/image_processing_idefics3.d.ts

@@ -0,0 +1,40 @@
+export class Idefics3ImageProcessor extends ImageProcessor {
+    constructor(config: any);
+    do_image_splitting: any;
+    max_image_size: any;
+    /**
+     * @typedef {import('../../utils/image.js').RawImage} RawImage
+     * @typedef {import('../../utils/tensor.js').Tensor} Tensor
+     */
+    /**
+     * Calculate size to resize images to, to be multiples of `vision_encoder_max_size` while preserving the aspect ratio.
+     * @param {Tensor} pixel_values Tensor of the image to resize.
+     * @param {number} vision_encoder_max_size Maximum size of the output image. If the image is larger than this size,
+     * it will be split into patches of this size, and the original image will be concatenated with the patches, resized to max_size.
+     */
+    get_resize_for_vision_encoder(pixel_values: import("../../utils/tensor.js").Tensor, vision_encoder_max_size: number): {
+        height: number;
+        width: number;
+    };
+    /** @param {RawImage|RawImage[]|RawImage[][]} images */
+    _call(images: import("../../utils/image.js").RawImage | import("../../utils/image.js").RawImage[] | import("../../utils/image.js").RawImage[][], { do_image_splitting, return_row_col_info, }?: {
+        do_image_splitting?: any;
+        return_row_col_info?: boolean;
+    }): Promise<{
+        rows?: any[][];
+        cols?: any[][];
+        pixel_values: import("../../utils/tensor.js").Tensor;
+        pixel_attention_mask: import("../../utils/tensor.js").Tensor;
+        original_sizes: import("../../base/image_processors_utils.js").HeightWidth[];
+        reshaped_input_sizes: import("../../base/image_processors_utils.js").HeightWidth[];
+    }>;
+    split_image(pixel_values: any, { longest_edge }: {
+        longest_edge: any;
+    }): Promise<{
+        frames: any[];
+        num_splits_h: number;
+        num_splits_w: number;
+    }>;
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_idefics3.d.ts.map

package/tfjs-types/models/idefics3/processing_idefics3.d.ts

@@ -0,0 +1,19 @@
+export class Idefics3Processor extends Processor {
+    static image_processor_class: typeof AutoImageProcessor;
+    static tokenizer_class: typeof AutoTokenizer;
+    fake_image_token: string;
+    image_token: string;
+    global_img_token: string;
+    /**
+     *
+     * @param {string|string[]} text
+     * @param {RawImage|RawImage[]|RawImage[][]} images
+     * @returns {Promise<any>}
+     */
+    _call(text: string | string[], images?: RawImage | RawImage[] | RawImage[][], options?: {}): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { RawImage } from "../../utils/image.js";
+import { AutoImageProcessor } from "../auto/image_processing_auto.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+//# sourceMappingURL=processing_idefics3.d.ts.map

package/tfjs-types/models/image_processors.d.ts

@@ -0,0 +1,44 @@
+export * from "./beit/image_processing_beit.js";
+export * from "./bit/image_processing_bit.js";
+export * from "./chinese_clip/image_processing_chinese_clip.js";
+export * from "./clip/image_processing_clip.js";
+export * from "./convnext/image_processing_convnext.js";
+export * from "./deit/image_processing_deit.js";
+export * from "./detr/image_processing_detr.js";
+export * from "./dinov3_vit/image_processing_dinov3_vit.js";
+export * from "./donut/image_processing_donut.js";
+export * from "./dpt/image_processing_dpt.js";
+export * from "./efficientnet/image_processing_efficientnet.js";
+export * from "./glpn/image_processing_glpn.js";
+export * from "./grounding_dino/image_processing_grounding_dino.js";
+export * from "./idefics3/image_processing_idefics3.js";
+export * from "./janus/image_processing_janus.js";
+export * from "./jina_clip/image_processing_jina_clip.js";
+export * from "./llava_onevision/image_processing_llava_onevision.js";
+export * from "./mask2former/image_processing_mask2former.js";
+export * from "./maskformer/image_processing_maskformer.js";
+export * from "./mobilenet_v1/image_processing_mobilenet_v1.js";
+export * from "./mobilenet_v2/image_processing_mobilenet_v2.js";
+export * from "./mobilenet_v3/image_processing_mobilenet_v3.js";
+export * from "./mobilenet_v4/image_processing_mobilenet_v4.js";
+export * from "./mobilevit/image_processing_mobilevit.js";
+export * from "./nougat/image_processing_nougat.js";
+export * from "./owlv2/image_processing_owlv2.js";
+export * from "./owlvit/image_processing_owlvit.js";
+export * from "./phi3_v/image_processing_phi3_v.js";
+export * from "./pixtral/image_processing_pixtral.js";
+export * from "./pvt/image_processing_pvt.js";
+export * from "./qwen2_vl/image_processing_qwen2_vl.js";
+export * from "./rt_detr/image_processing_rt_detr.js";
+export * from "./sam/image_processing_sam.js";
+export * from "./sam2/image_processing_sam2.js";
+export * from "./sam3/image_processing_sam3.js";
+export * from "./segformer/image_processing_segformer.js";
+export * from "./siglip/image_processing_siglip.js";
+export * from "./smolvlm/image_processing_smolvlm.js";
+export * from "./swin2sr/image_processing_swin2sr.js";
+export * from "./vit/image_processing_vit.js";
+export * from "./vitmatte/image_processing_vitmatte.js";
+export * from "./vitpose/image_processing_vitpose.js";
+export * from "./yolos/image_processing_yolos.js";
+//# sourceMappingURL=image_processors.d.ts.map

package/tfjs-types/models/janus/image_processing_janus.d.ts

@@ -0,0 +1,7 @@
+export class VLMImageProcessor extends ImageProcessor {
+    constructor(config: any);
+    constant_values: any;
+    pad_image(pixelData: any, imgDims: any, padSize: any, options: any): [Float32Array<ArrayBufferLike>, number[]];
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_janus.d.ts.map

package/tfjs-types/models/janus/processing_janus.d.ts

@@ -0,0 +1,77 @@
+export class VLChatProcessor extends Processor {
+    static image_processor_class: typeof AutoImageProcessor;
+    static tokenizer_class: typeof AutoTokenizer;
+    constructor(config: any, components: any, chat_template: any);
+    image_tag: any;
+    image_start_tag: any;
+    image_end_tag: any;
+    num_image_tokens: any;
+    /**
+     * @typedef {Object} MultimodalMessageProperties Additional properties for multimodal messages.
+     * @property {(RawImage | string | URL)[]} [images] The images in the message.
+     * @typedef {(import('../../tokenizers.js').Message & MultimodalMessageProperties)[]} MultimodalConversation The conversation possibly containing multimodal inputs.
+     */
+    /**
+     * @typedef {Object} VLCChatProcessorResult The processed input.
+     * @property {Tensor} input_ids The input IDs.
+     * @property {Tensor} attention_mask The attention mask.
+     * @property {Tensor} images_seq_mask The image sequence mask.
+     * @property {Tensor} images_emb_mask The image embedding mask.
+     */
+    /**
+     * @param {MultimodalConversation} conversation The chat messages to process.
+     * @param {Object} options Additional options for processing.
+     * @param {RawImage|RawImage[]} [options.images] The images to process, if not set in the conversation.
+     * @param {string} [options.chat_template="default"] The chat template to use.
+     * @returns {Promise<VLCChatProcessorResult | VLCChatProcessorResult & import('../../base/image_processors_utils.js').ImageProcessorResult>} The processed input.
+     */
+    _call(conversation: (import("../../tokenizers.js").Message & {
+        /**
+         * The images in the message.
+         */
+        images?: (RawImage | string | URL)[];
+    })[], { images, chat_template, }?: {
+        images?: RawImage | RawImage[];
+        chat_template?: string;
+    }): Promise<{
+        /**
+         * The input IDs.
+         */
+        input_ids: Tensor;
+        /**
+         * The attention mask.
+         */
+        attention_mask: Tensor;
+        /**
+         * The image sequence mask.
+         */
+        images_seq_mask: Tensor;
+        /**
+         * The image embedding mask.
+         */
+        images_emb_mask: Tensor;
+    } | ({
+        /**
+         * The input IDs.
+         */
+        input_ids: Tensor;
+        /**
+         * The attention mask.
+         */
+        attention_mask: Tensor;
+        /**
+         * The image sequence mask.
+         */
+        images_seq_mask: Tensor;
+        /**
+         * The image embedding mask.
+         */
+        images_emb_mask: Tensor;
+    } & import("../../base/image_processors_utils.js").ImageProcessorResult)>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { RawImage } from "../../utils/image.js";
+import { Tensor } from "../../utils/tensor.js";
+import { AutoImageProcessor } from "../auto/image_processing_auto.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+//# sourceMappingURL=processing_janus.d.ts.map

package/tfjs-types/models/jina_clip/image_processing_jina_clip.d.ts

@@ -0,0 +1,5 @@
+export class JinaCLIPImageProcessor extends ImageProcessor {
+    constructor(config: any);
+}
+import { ImageProcessor } from "../../base/image_processors_utils.js";
+//# sourceMappingURL=image_processing_jina_clip.d.ts.map

package/tfjs-types/models/jina_clip/processing_jina_clip.d.ts

@@ -0,0 +1,9 @@
+export class JinaCLIPProcessor extends Processor {
+    static tokenizer_class: typeof AutoTokenizer;
+    static image_processor_class: typeof AutoImageProcessor;
+    _call(text?: any, images?: any, kwargs?: {}): Promise<any>;
+}
+import { Processor } from "../../base/processing_utils.js";
+import { AutoTokenizer } from "../../tokenizers.js";
+import { AutoImageProcessor } from "../auto/image_processing_auto.js";
+//# sourceMappingURL=processing_jina_clip.d.ts.map