@huggingface/transformers 3.4.0 → 3.4.2

package/dist/transformers.node.cjs

@@ -3684,6 +3684,7 @@ function getNormalizedConfig(config) {
         // Sub-configs
         case 'llava':
         case 'paligemma':
+        case 'gemma3':
         case 'florence2':
         case 'llava_onevision':
         case 'idefics3':
@@ -3743,6 +3744,7 @@ function getNormalizedConfig(config) {
             break;
         case 'gemma':
         case 'gemma2':
+        case 'gemma3_text':
         case 'glm':
         case 'helium':
             mapping['num_heads'] = 'num_key_value_heads';
@@ -4074,7 +4076,7 @@ __webpack_require__.r(__webpack_exports__);
 
 
 
-const VERSION = '3.4.0';
+const VERSION = '3.4.2';
 
 // Check if various APIs are available (depends on environment)
 const IS_BROWSER_ENV = typeof window !== "undefined" && typeof window.document !== "undefined";
@@ -6218,6 +6220,9 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony export */   Gemma2ForCausalLM: () => (/* binding */ Gemma2ForCausalLM),
 /* harmony export */   Gemma2Model: () => (/* binding */ Gemma2Model),
 /* harmony export */   Gemma2PreTrainedModel: () => (/* binding */ Gemma2PreTrainedModel),
+/* harmony export */   Gemma3ForCausalLM: () => (/* binding */ Gemma3ForCausalLM),
+/* harmony export */   Gemma3Model: () => (/* binding */ Gemma3Model),
+/* harmony export */   Gemma3PreTrainedModel: () => (/* binding */ Gemma3PreTrainedModel),
 /* harmony export */   GemmaForCausalLM: () => (/* binding */ GemmaForCausalLM),
 /* harmony export */   GemmaModel: () => (/* binding */ GemmaModel),
 /* harmony export */   GemmaPreTrainedModel: () => (/* binding */ GemmaPreTrainedModel),
@@ -6288,6 +6293,10 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony export */   MaskFormerModel: () => (/* binding */ MaskFormerModel),
 /* harmony export */   MaskFormerPreTrainedModel: () => (/* binding */ MaskFormerPreTrainedModel),
 /* harmony export */   MaskedLMOutput: () => (/* binding */ MaskedLMOutput),
+/* harmony export */   Metric3DForDepthEstimation: () => (/* binding */ Metric3DForDepthEstimation),
+/* harmony export */   Metric3DPreTrainedModel: () => (/* binding */ Metric3DPreTrainedModel),
+/* harmony export */   Metric3Dv2ForDepthEstimation: () => (/* binding */ Metric3Dv2ForDepthEstimation),
+/* harmony export */   Metric3Dv2PreTrainedModel: () => (/* binding */ Metric3Dv2PreTrainedModel),
 /* harmony export */   MgpstrForSceneTextRecognition: () => (/* binding */ MgpstrForSceneTextRecognition),
 /* harmony export */   MgpstrModelOutput: () => (/* binding */ MgpstrModelOutput),
 /* harmony export */   MgpstrPreTrainedModel: () => (/* binding */ MgpstrPreTrainedModel),
@@ -6399,10 +6408,18 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony export */   Qwen2PreTrainedModel: () => (/* binding */ Qwen2PreTrainedModel),
 /* harmony export */   Qwen2VLForConditionalGeneration: () => (/* binding */ Qwen2VLForConditionalGeneration),
 /* harmony export */   Qwen2VLPreTrainedModel: () => (/* binding */ Qwen2VLPreTrainedModel),
+/* harmony export */   RFDetrForObjectDetection: () => (/* binding */ RFDetrForObjectDetection),
+/* harmony export */   RFDetrModel: () => (/* binding */ RFDetrModel),
+/* harmony export */   RFDetrObjectDetectionOutput: () => (/* binding */ RFDetrObjectDetectionOutput),
+/* harmony export */   RFDetrPreTrainedModel: () => (/* binding */ RFDetrPreTrainedModel),
 /* harmony export */   RTDetrForObjectDetection: () => (/* binding */ RTDetrForObjectDetection),
 /* harmony export */   RTDetrModel: () => (/* binding */ RTDetrModel),
 /* harmony export */   RTDetrObjectDetectionOutput: () => (/* binding */ RTDetrObjectDetectionOutput),
 /* harmony export */   RTDetrPreTrainedModel: () => (/* binding */ RTDetrPreTrainedModel),
+/* harmony export */   RTDetrV2ForObjectDetection: () => (/* binding */ RTDetrV2ForObjectDetection),
+/* harmony export */   RTDetrV2Model: () => (/* binding */ RTDetrV2Model),
+/* harmony export */   RTDetrV2ObjectDetectionOutput: () => (/* binding */ RTDetrV2ObjectDetectionOutput),
+/* harmony export */   RTDetrV2PreTrainedModel: () => (/* binding */ RTDetrV2PreTrainedModel),
 /* harmony export */   ResNetForImageClassification: () => (/* binding */ ResNetForImageClassification),
 /* harmony export */   ResNetModel: () => (/* binding */ ResNetModel),
 /* harmony export */   ResNetPreTrainedModel: () => (/* binding */ ResNetPreTrainedModel),
@@ -6436,6 +6453,10 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony export */   SiglipTextModel: () => (/* binding */ SiglipTextModel),
 /* harmony export */   SiglipVisionModel: () => (/* binding */ SiglipVisionModel),
 /* harmony export */   SmolVLMForConditionalGeneration: () => (/* binding */ SmolVLMForConditionalGeneration),
+/* harmony export */   SnacDecoderModel: () => (/* binding */ SnacDecoderModel),
+/* harmony export */   SnacEncoderModel: () => (/* binding */ SnacEncoderModel),
+/* harmony export */   SnacModel: () => (/* binding */ SnacModel),
+/* harmony export */   SnacPreTrainedModel: () => (/* binding */ SnacPreTrainedModel),
 /* harmony export */   SpeechT5ForSpeechToText: () => (/* binding */ SpeechT5ForSpeechToText),
 /* harmony export */   SpeechT5ForTextToSpeech: () => (/* binding */ SpeechT5ForTextToSpeech),
 /* harmony export */   SpeechT5HifiGan: () => (/* binding */ SpeechT5HifiGan),
@@ -7094,8 +7115,8 @@ async function decoderForward(self, model_inputs, is_encoder_decoder = false) {
         new_model_inputs.use_cache_branch = boolTensor(!!past_key_values);
     }
     if (session.inputNames.includes('position_ids') && new_model_inputs.attention_mask && !new_model_inputs.position_ids) {
-        // NOTE: Handle a special case for paligemma models, where positions are 1-indexed
-        const start_index = self.config.model_type === 'paligemma' ? 1 : 0;
+        // NOTE: Handle a special case for paligemma/gemma3 models, where positions are 1-indexed
+        const start_index = ['paligemma', 'gemma3_text', 'gemma3'].includes(self.config.model_type) ? 1 : 0;
         new_model_inputs.position_ids = createPositionIds(new_model_inputs, past_key_values, start_index);
     }
 
@@ -11020,6 +11041,23 @@ class Gemma2Model extends Gemma2PreTrainedModel { }
 class Gemma2ForCausalLM extends Gemma2PreTrainedModel { }
 //////////////////////////////////////////////////
 
+
+//////////////////////////////////////////////////
+// Gemma3 models
+
+/**
+ * The bare Gemma3 Model outputting raw hidden-states without any specific head on top.
+ */
+class Gemma3PreTrainedModel extends PreTrainedModel { }
+/**
+ * The bare Gemma3 Model outputting raw hidden-states without any specific head on top.
+ */
+class Gemma3Model extends Gemma3PreTrainedModel { }
+
+class Gemma3ForCausalLM extends Gemma3PreTrainedModel { }
+//////////////////////////////////////////////////
+
+
 //////////////////////////////////////////////////
 class OpenELMPreTrainedModel extends PreTrainedModel { }
 class OpenELMModel extends OpenELMPreTrainedModel { }
@@ -11664,6 +11702,37 @@ class RTDetrObjectDetectionOutput extends ModelOutput {
 }
 //////////////////////////////////////////////////
 
+
+//////////////////////////////////////////////////
+class RTDetrV2PreTrainedModel extends PreTrainedModel { }
+class RTDetrV2Model extends RTDetrV2PreTrainedModel { }
+class RTDetrV2ForObjectDetection extends RTDetrV2PreTrainedModel {
+    /**
+     * @param {any} model_inputs
+     */
+    async _call(model_inputs) {
+        return new RTDetrV2ObjectDetectionOutput(await super._call(model_inputs));
+    }
+}
+
+class RTDetrV2ObjectDetectionOutput extends RTDetrObjectDetectionOutput {}
+//////////////////////////////////////////////////
+
+//////////////////////////////////////////////////
+class RFDetrPreTrainedModel extends PreTrainedModel { }
+class RFDetrModel extends RFDetrPreTrainedModel { }
+class RFDetrForObjectDetection extends RFDetrPreTrainedModel {
+    /**
+     * @param {any} model_inputs
+     */
+    async _call(model_inputs) {
+        return new RFDetrObjectDetectionOutput(await super._call(model_inputs));
+    }
+}
+
+class RFDetrObjectDetectionOutput extends RTDetrObjectDetectionOutput {}
+//////////////////////////////////////////////////
+
 //////////////////////////////////////////////////
 class TableTransformerPreTrainedModel extends PreTrainedModel { }
 
@@ -11872,6 +11941,16 @@ class DepthProPreTrainedModel extends PreTrainedModel { }
 class DepthProForDepthEstimation extends DepthProPreTrainedModel { }
 //////////////////////////////////////////////////
 
+//////////////////////////////////////////////////
+class Metric3DPreTrainedModel extends PreTrainedModel { }
+class Metric3DForDepthEstimation extends Metric3DPreTrainedModel { }
+//////////////////////////////////////////////////
+
+//////////////////////////////////////////////////
+class Metric3Dv2PreTrainedModel extends PreTrainedModel { }
+class Metric3Dv2ForDepthEstimation extends Metric3Dv2PreTrainedModel { }
+//////////////////////////////////////////////////
+
 //////////////////////////////////////////////////
 class MaskFormerPreTrainedModel extends PreTrainedModel { }
 class MaskFormerModel extends MaskFormerPreTrainedModel { }
@@ -13787,6 +13866,60 @@ class DacDecoderModel extends DacPreTrainedModel {
 }
 //////////////////////////////////////////////////
 
+
+//////////////////////////////////////////////////
+// Snac models
+class SnacPreTrainedModel extends PreTrainedModel {
+    main_input_name = 'input_values';
+    forward_params = ['input_values'];
+}
+
+/**
+ * The SNAC (Multi-Scale Neural Audio Codec) model.
+ */
+class SnacModel extends SnacPreTrainedModel {
+    /**
+     * Encodes the input audio waveform into discrete codes.
+     * @param {Object} inputs Model inputs
+     * @param {Tensor} [inputs.input_values] Float values of the input audio waveform, of shape `(batch_size, channels, sequence_length)`).
+     * @returns {Promise<Record<string, Tensor>>} The output tensors of shape `(batch_size, num_codebooks, sequence_length)`.
+     */
+    async encode(inputs) {
+        return await sessionRun(this.sessions['encoder_model'], inputs);
+    }
+
+    /**
+     * Decodes the given frames into an output audio waveform.
+     * @param {Record<string, Tensor>} inputs The encoded audio codes.
+     * @returns {Promise<{audio_values: Tensor}>} The output tensor of shape `(batch_size, num_channels, sequence_length)`.
+     */
+    async decode(inputs) {
+        return await sessionRun(this.sessions['decoder_model'], inputs);
+    }
+}
+
+class SnacEncoderModel extends SnacPreTrainedModel {
+    /** @type {typeof PreTrainedModel.from_pretrained} */
+    static async from_pretrained(pretrained_model_name_or_path, options = {}) {
+        return super.from_pretrained(pretrained_model_name_or_path, {
+            ...options,
+            // Update default model file name if not provided
+            model_file_name: options.model_file_name ?? 'encoder_model',
+        });
+    }
+}
+class SnacDecoderModel extends SnacPreTrainedModel {
+    /** @type {typeof PreTrainedModel.from_pretrained} */
+    static async from_pretrained(pretrained_model_name_or_path, options = {}) {
+        return super.from_pretrained(pretrained_model_name_or_path, {
+            ...options,
+            // Update default model file name if not provided
+            model_file_name: options.model_file_name ?? 'decoder_model',
+        });
+    }
+}
+//////////////////////////////////////////////////
+
 //////////////////////////////////////////////////
 // AutoModels, used to simplify construction of PreTrainedModels
 // (uses config to instantiate correct class)
@@ -13907,6 +14040,8 @@ const MODEL_MAPPING_NAMES_ENCODER_ONLY = new Map([
 
     ['detr', ['DetrModel', DetrModel]],
     ['rt_detr', ['RTDetrModel', RTDetrModel]],
+    ['rt_detr_v2', ['RTDetrV2Model', RTDetrV2Model]],
+    ['rf_detr', ['RFDetrModel', RFDetrModel]],
     ['table-transformer', ['TableTransformerModel', TableTransformerModel]],
     ['vit', ['ViTModel', ViTModel]],
     ['ijepa', ['IJepaModel', IJepaModel]],
@@ -13968,6 +14103,7 @@ const MODEL_MAPPING_NAMES_ENCODER_DECODER = new Map([
 const MODEL_MAPPING_NAMES_AUTO_ENCODER = new Map([
     ['mimi', ['MimiModel', MimiModel]],
     ['dac', ['DacModel', DacModel]],
+    ['snac', ['SnacModel', SnacModel]],
 ]);
 
 const MODEL_MAPPING_NAMES_DECODER_ONLY = new Map([
@@ -13988,6 +14124,7 @@ const MODEL_MAPPING_NAMES_DECODER_ONLY = new Map([
     ['cohere', ['CohereModel', CohereModel]],
     ['gemma', ['GemmaModel', GemmaModel]],
     ['gemma2', ['Gemma2Model', Gemma2Model]],
+    ['gemma3_text', ['Gemma3Model', Gemma3Model]],
     ['helium', ['HeliumModel', HeliumModel]],
     ['glm', ['GlmModel', GlmModel]],
     ['openelm', ['OpenELMModel', OpenELMModel]],
@@ -14087,6 +14224,7 @@ const MODEL_FOR_CAUSAL_LM_MAPPING_NAMES = new Map([
     ['cohere', ['CohereForCausalLM', CohereForCausalLM]],
     ['gemma', ['GemmaForCausalLM', GemmaForCausalLM]],
     ['gemma2', ['Gemma2ForCausalLM', Gemma2ForCausalLM]],
+    ['gemma3_text', ['Gemma3ForCausalLM', Gemma3ForCausalLM]],
     ['helium', ['HeliumForCausalLM', HeliumForCausalLM]],
     ['glm', ['GlmForCausalLM', GlmForCausalLM]],
     ['openelm', ['OpenELMForCausalLM', OpenELMForCausalLM]],
@@ -14203,6 +14341,8 @@ const MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES = new Map([
 const MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES = new Map([
     ['detr', ['DetrForObjectDetection', DetrForObjectDetection]],
     ['rt_detr', ['RTDetrForObjectDetection', RTDetrForObjectDetection]],
+    ['rt_detr_v2', ['RTDetrV2ForObjectDetection', RTDetrV2ForObjectDetection]],
+    ['rf_detr', ['RFDetrForObjectDetection', RFDetrForObjectDetection]],
     ['table-transformer', ['TableTransformerForObjectDetection', TableTransformerForObjectDetection]],
     ['yolos', ['YolosForObjectDetection', YolosForObjectDetection]],
 ]);
@@ -14288,6 +14428,8 @@ const MODEL_FOR_DEPTH_ESTIMATION_MAPPING_NAMES = new Map([
     ['glpn', ['GLPNForDepthEstimation', GLPNForDepthEstimation]],
     ['sapiens', ['SapiensForDepthEstimation', SapiensForDepthEstimation]],
     ['depth_pro', ['DepthProForDepthEstimation', DepthProForDepthEstimation]],
+    ['metric3d', ['Metric3DForDepthEstimation', Metric3DForDepthEstimation]],
+    ['metric3dv2', ['Metric3Dv2ForDepthEstimation', Metric3Dv2ForDepthEstimation]],
 ])
 
 const MODEL_FOR_NORMAL_ESTIMATION_MAPPING_NAMES = new Map([
@@ -14373,6 +14515,8 @@ const CUSTOM_MAPPING = [
     ['DacDecoderModel', DacDecoderModel, MODEL_TYPES.EncoderOnly],
     ['MimiEncoderModel', MimiEncoderModel, MODEL_TYPES.EncoderOnly],
     ['MimiDecoderModel', MimiDecoderModel, MODEL_TYPES.EncoderOnly],
+    ['SnacEncoderModel', SnacEncoderModel, MODEL_TYPES.EncoderOnly],
+    ['SnacDecoderModel', SnacDecoderModel, MODEL_TYPES.EncoderOnly],
 ]
 for (const [name, model, type] of CUSTOM_MAPPING) {
     MODEL_TYPE_MAPPING.set(name, type);
@@ -15689,14 +15833,15 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony export */   ClapFeatureExtractor: () => (/* reexport safe */ _clap_feature_extraction_clap_js__WEBPACK_IMPORTED_MODULE_2__.ClapFeatureExtractor),
 /* harmony export */   DacFeatureExtractor: () => (/* reexport safe */ _dac_feature_extraction_dac_js__WEBPACK_IMPORTED_MODULE_3__.DacFeatureExtractor),
 /* harmony export */   EncodecFeatureExtractor: () => (/* reexport safe */ _encodec_feature_extraction_encodec_js__WEBPACK_IMPORTED_MODULE_1__.EncodecFeatureExtractor),
-/* harmony export */   ImageFeatureExtractor: () => (/* reexport safe */ _base_image_processors_utils_js__WEBPACK_IMPORTED_MODULE_11__.ImageProcessor),
+/* harmony export */   ImageFeatureExtractor: () => (/* reexport safe */ _base_image_processors_utils_js__WEBPACK_IMPORTED_MODULE_12__.ImageProcessor),
 /* harmony export */   MoonshineFeatureExtractor: () => (/* reexport safe */ _moonshine_feature_extraction_moonshine_js__WEBPACK_IMPORTED_MODULE_4__.MoonshineFeatureExtractor),
 /* harmony export */   PyAnnoteFeatureExtractor: () => (/* reexport safe */ _pyannote_feature_extraction_pyannote_js__WEBPACK_IMPORTED_MODULE_5__.PyAnnoteFeatureExtractor),
 /* harmony export */   SeamlessM4TFeatureExtractor: () => (/* reexport safe */ _seamless_m4t_feature_extraction_seamless_m4t_js__WEBPACK_IMPORTED_MODULE_6__.SeamlessM4TFeatureExtractor),
-/* harmony export */   SpeechT5FeatureExtractor: () => (/* reexport safe */ _speecht5_feature_extraction_speecht5_js__WEBPACK_IMPORTED_MODULE_7__.SpeechT5FeatureExtractor),
-/* harmony export */   Wav2Vec2FeatureExtractor: () => (/* reexport safe */ _wav2vec2_feature_extraction_wav2vec2_js__WEBPACK_IMPORTED_MODULE_8__.Wav2Vec2FeatureExtractor),
-/* harmony export */   WeSpeakerFeatureExtractor: () => (/* reexport safe */ _wespeaker_feature_extraction_wespeaker_js__WEBPACK_IMPORTED_MODULE_9__.WeSpeakerFeatureExtractor),
-/* harmony export */   WhisperFeatureExtractor: () => (/* reexport safe */ _whisper_feature_extraction_whisper_js__WEBPACK_IMPORTED_MODULE_10__.WhisperFeatureExtractor)
+/* harmony export */   SnacFeatureExtractor: () => (/* reexport safe */ _snac_feature_extraction_snac_js__WEBPACK_IMPORTED_MODULE_7__.SnacFeatureExtractor),
+/* harmony export */   SpeechT5FeatureExtractor: () => (/* reexport safe */ _speecht5_feature_extraction_speecht5_js__WEBPACK_IMPORTED_MODULE_8__.SpeechT5FeatureExtractor),
+/* harmony export */   Wav2Vec2FeatureExtractor: () => (/* reexport safe */ _wav2vec2_feature_extraction_wav2vec2_js__WEBPACK_IMPORTED_MODULE_9__.Wav2Vec2FeatureExtractor),
+/* harmony export */   WeSpeakerFeatureExtractor: () => (/* reexport safe */ _wespeaker_feature_extraction_wespeaker_js__WEBPACK_IMPORTED_MODULE_10__.WeSpeakerFeatureExtractor),
+/* harmony export */   WhisperFeatureExtractor: () => (/* reexport safe */ _whisper_feature_extraction_whisper_js__WEBPACK_IMPORTED_MODULE_11__.WhisperFeatureExtractor)
 /* harmony export */ });
 /* harmony import */ var _audio_spectrogram_transformer_feature_extraction_audio_spectrogram_transformer_js__WEBPACK_IMPORTED_MODULE_0__ = __webpack_require__(/*! ./audio_spectrogram_transformer/feature_extraction_audio_spectrogram_transformer.js */ "./src/models/audio_spectrogram_transformer/feature_extraction_audio_spectrogram_transformer.js");
 /* harmony import */ var _encodec_feature_extraction_encodec_js__WEBPACK_IMPORTED_MODULE_1__ = __webpack_require__(/*! ./encodec/feature_extraction_encodec.js */ "./src/models/encodec/feature_extraction_encodec.js");
@@ -15705,11 +15850,13 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony import */ var _moonshine_feature_extraction_moonshine_js__WEBPACK_IMPORTED_MODULE_4__ = __webpack_require__(/*! ./moonshine/feature_extraction_moonshine.js */ "./src/models/moonshine/feature_extraction_moonshine.js");
 /* harmony import */ var _pyannote_feature_extraction_pyannote_js__WEBPACK_IMPORTED_MODULE_5__ = __webpack_require__(/*! ./pyannote/feature_extraction_pyannote.js */ "./src/models/pyannote/feature_extraction_pyannote.js");
 /* harmony import */ var _seamless_m4t_feature_extraction_seamless_m4t_js__WEBPACK_IMPORTED_MODULE_6__ = __webpack_require__(/*! ./seamless_m4t/feature_extraction_seamless_m4t.js */ "./src/models/seamless_m4t/feature_extraction_seamless_m4t.js");
-/* harmony import */ var _speecht5_feature_extraction_speecht5_js__WEBPACK_IMPORTED_MODULE_7__ = __webpack_require__(/*! ./speecht5/feature_extraction_speecht5.js */ "./src/models/speecht5/feature_extraction_speecht5.js");
-/* harmony import */ var _wav2vec2_feature_extraction_wav2vec2_js__WEBPACK_IMPORTED_MODULE_8__ = __webpack_require__(/*! ./wav2vec2/feature_extraction_wav2vec2.js */ "./src/models/wav2vec2/feature_extraction_wav2vec2.js");
-/* harmony import */ var _wespeaker_feature_extraction_wespeaker_js__WEBPACK_IMPORTED_MODULE_9__ = __webpack_require__(/*! ./wespeaker/feature_extraction_wespeaker.js */ "./src/models/wespeaker/feature_extraction_wespeaker.js");
-/* harmony import */ var _whisper_feature_extraction_whisper_js__WEBPACK_IMPORTED_MODULE_10__ = __webpack_require__(/*! ./whisper/feature_extraction_whisper.js */ "./src/models/whisper/feature_extraction_whisper.js");
-/* harmony import */ var _base_image_processors_utils_js__WEBPACK_IMPORTED_MODULE_11__ = __webpack_require__(/*! ../base/image_processors_utils.js */ "./src/base/image_processors_utils.js");
+/* harmony import */ var _snac_feature_extraction_snac_js__WEBPACK_IMPORTED_MODULE_7__ = __webpack_require__(/*! ./snac/feature_extraction_snac.js */ "./src/models/snac/feature_extraction_snac.js");
+/* harmony import */ var _speecht5_feature_extraction_speecht5_js__WEBPACK_IMPORTED_MODULE_8__ = __webpack_require__(/*! ./speecht5/feature_extraction_speecht5.js */ "./src/models/speecht5/feature_extraction_speecht5.js");
+/* harmony import */ var _wav2vec2_feature_extraction_wav2vec2_js__WEBPACK_IMPORTED_MODULE_9__ = __webpack_require__(/*! ./wav2vec2/feature_extraction_wav2vec2.js */ "./src/models/wav2vec2/feature_extraction_wav2vec2.js");
+/* harmony import */ var _wespeaker_feature_extraction_wespeaker_js__WEBPACK_IMPORTED_MODULE_10__ = __webpack_require__(/*! ./wespeaker/feature_extraction_wespeaker.js */ "./src/models/wespeaker/feature_extraction_wespeaker.js");
+/* harmony import */ var _whisper_feature_extraction_whisper_js__WEBPACK_IMPORTED_MODULE_11__ = __webpack_require__(/*! ./whisper/feature_extraction_whisper.js */ "./src/models/whisper/feature_extraction_whisper.js");
+/* harmony import */ var _base_image_processors_utils_js__WEBPACK_IMPORTED_MODULE_12__ = __webpack_require__(/*! ../base/image_processors_utils.js */ "./src/base/image_processors_utils.js");
+
 
 
 
@@ -18742,6 +18889,25 @@ __webpack_require__.r(__webpack_exports__);
 
 
 
+/***/ }),
+
+/***/ "./src/models/snac/feature_extraction_snac.js":
+/*!****************************************************!*\
+  !*** ./src/models/snac/feature_extraction_snac.js ***!
+  \****************************************************/
+/***/ ((__unused_webpack___webpack_module__, __webpack_exports__, __webpack_require__) => {
+
+"use strict";
+__webpack_require__.r(__webpack_exports__);
+/* harmony export */ __webpack_require__.d(__webpack_exports__, {
+/* harmony export */   SnacFeatureExtractor: () => (/* binding */ SnacFeatureExtractor)
+/* harmony export */ });
+/* harmony import */ var _dac_feature_extraction_dac_js__WEBPACK_IMPORTED_MODULE_0__ = __webpack_require__(/*! ../dac/feature_extraction_dac.js */ "./src/models/dac/feature_extraction_dac.js");
+
+
+class SnacFeatureExtractor extends _dac_feature_extraction_dac_js__WEBPACK_IMPORTED_MODULE_0__.DacFeatureExtractor { }
+
+
 /***/ }),
 
 /***/ "./src/models/speecht5/feature_extraction_speecht5.js":
@@ -19999,16 +20165,16 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony import */ var _utils_image_js__WEBPACK_IMPORTED_MODULE_9__ = __webpack_require__(/*! ./utils/image.js */ "./src/utils/image.js");
 /**
  * @file Pipelines provide a high-level, easy to use, API for running machine learning models.
- * 
+ *
  * **Example:** Instantiate pipeline using the `pipeline` function.
  * ```javascript
  * import { pipeline } from '@huggingface/transformers';
- * 
+ *
  * const classifier = await pipeline('sentiment-analysis');
  * const output = await classifier('I love transformers!');
  * // [{'label': 'POSITIVE', 'score': 0.999817686}]
  * ```
- * 
+ *
  * @module pipelines
  */
 
@@ -20027,7 +20193,7 @@ __webpack_require__.r(__webpack_exports__);
 
 
 /**
- * @typedef {string | RawImage | URL} ImageInput
+ * @typedef {string | RawImage | URL | Blob | HTMLCanvasElement | OffscreenCanvas} ImageInput
  * @typedef {ImageInput|ImageInput[]} ImagePipelineInputs
  */
 
@@ -20101,7 +20267,7 @@ function get_bounding_box(box, asInteger) {
 /**
  * @callback DisposeType Disposes the item.
  * @returns {Promise<void>} A promise that resolves when the item has been disposed.
- * 
+ *
  * @typedef {Object} Disposable
  * @property {DisposeType} dispose A promise that resolves when the pipeline has been disposed.
  */
@@ -20138,7 +20304,7 @@ class Pipeline extends _utils_generic_js__WEBPACK_IMPORTED_MODULE_4__.Callable {
  * @property {string} task The task of the pipeline. Useful for specifying subtasks.
  * @property {PreTrainedModel} model The model used by the pipeline.
  * @property {PreTrainedTokenizer} tokenizer The tokenizer used by the pipeline.
- * 
+ *
  * @typedef {ModelTokenizerConstructorArgs} TextPipelineConstructorArgs An object used to instantiate a text-based pipeline.
  */
 
@@ -20147,7 +20313,7 @@ class Pipeline extends _utils_generic_js__WEBPACK_IMPORTED_MODULE_4__.Callable {
  * @property {string} task The task of the pipeline. Useful for specifying subtasks.
  * @property {PreTrainedModel} model The model used by the pipeline.
  * @property {Processor} processor The processor used by the pipeline.
- * 
+ *
  * @typedef {ModelProcessorConstructorArgs} AudioPipelineConstructorArgs An object used to instantiate an audio-based pipeline.
  * @typedef {ModelProcessorConstructorArgs} ImagePipelineConstructorArgs An object used to instantiate an image-based pipeline.
  */
@@ -20159,7 +20325,7 @@ class Pipeline extends _utils_generic_js__WEBPACK_IMPORTED_MODULE_4__.Callable {
  * @property {PreTrainedModel} model The model used by the pipeline.
  * @property {PreTrainedTokenizer} tokenizer The tokenizer used by the pipeline.
  * @property {Processor} processor The processor used by the pipeline.
- * 
+ *
  * @typedef {ModelTokenizerProcessorConstructorArgs} TextAudioPipelineConstructorArgs An object used to instantiate a text- and audio-based pipeline.
  * @typedef {ModelTokenizerProcessorConstructorArgs} TextImagePipelineConstructorArgs An object used to instantiate a text- and image-based pipeline.
  */
@@ -20169,15 +20335,15 @@ class Pipeline extends _utils_generic_js__WEBPACK_IMPORTED_MODULE_4__.Callable {
  * @property {string} label The label predicted.
  * @property {number} score The corresponding probability.
  * @typedef {TextClassificationSingle[]} TextClassificationOutput
- * 
+ *
  * @typedef {Object} TextClassificationPipelineOptions Parameters specific to text classification pipelines.
  * @property {number} [top_k=1] The number of top predictions to be returned.
- * 
+ *
  * @callback TextClassificationPipelineCallback Classify the text(s) given as inputs.
  * @param {string|string[]} texts The input text(s) to be classified.
  * @param {TextClassificationPipelineOptions} [options] The options to use for text classification.
  * @returns {Promise<TextClassificationOutput|TextClassificationOutput[]>} An array or object containing the predicted labels and scores.
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & TextClassificationPipelineCallback & Disposable} TextClassificationPipelineType
  */
 
@@ -20190,7 +20356,7 @@ class Pipeline extends _utils_generic_js__WEBPACK_IMPORTED_MODULE_4__.Callable {
  * const output = await classifier('I love transformers!');
  * // [{ label: 'POSITIVE', score: 0.999788761138916 }]
  * ```
- * 
+ *
  * **Example:** Multilingual sentiment-analysis w/ `Xenova/bert-base-multilingual-uncased-sentiment` (and return top 5 classes).
  * ```javascript
  * const classifier = await pipeline('sentiment-analysis', 'Xenova/bert-base-multilingual-uncased-sentiment');
@@ -20203,7 +20369,7 @@ class Pipeline extends _utils_generic_js__WEBPACK_IMPORTED_MODULE_4__.Callable {
  * //   { label: '2 stars', score: 0.0009423971059732139 }
  * // ]
  * ```
- * 
+ *
  * **Example:** Toxic comment classification w/ `Xenova/toxic-bert` (and return all classes).
  * ```javascript
  * const classifier = await pipeline('text-classification', 'Xenova/toxic-bert');
@@ -20288,21 +20454,21 @@ class TextClassificationPipeline extends (/** @type {new (options: TextPipelineC
  * @property {number} [start] The index of the start of the corresponding entity in the sentence.
  * @property {number} [end] The index of the end of the corresponding entity in the sentence.
  * @typedef {TokenClassificationSingle[]} TokenClassificationOutput
- * 
+ *
  * @typedef {Object} TokenClassificationPipelineOptions Parameters specific to token classification pipelines.
  * @property {string[]} [ignore_labels] A list of labels to ignore.
- * 
+ *
  * @callback TokenClassificationPipelineCallback Classify each token of the text(s) given as inputs.
  * @param {string|string[]} texts One or several texts (or one list of texts) for token classification.
  * @param {TokenClassificationPipelineOptions} [options] The options to use for token classification.
  * @returns {Promise<TokenClassificationOutput|TokenClassificationOutput[]>} The result.
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & TokenClassificationPipelineCallback & Disposable} TokenClassificationPipelineType
  */
 
 /**
  * Named Entity Recognition pipeline using any `ModelForTokenClassification`.
- * 
+ *
  * **Example:** Perform named entity recognition with `Xenova/bert-base-NER`.
  * ```javascript
  * const classifier = await pipeline('token-classification', 'Xenova/bert-base-NER');
@@ -20312,7 +20478,7 @@ class TextClassificationPipeline extends (/** @type {new (options: TextPipelineC
  * //   { entity: 'B-LOC', score: 0.9994474053382874, index: 9, word: 'London' }
  * // ]
  * ```
- * 
+ *
  * **Example:** Perform named entity recognition with `Xenova/bert-base-NER` (and return all labels).
  * ```javascript
  * const classifier = await pipeline('token-classification', 'Xenova/bert-base-NER');
@@ -20408,22 +20574,22 @@ class TokenClassificationPipeline extends (/** @type {new (options: TextPipeline
  * @property {number} [start] The character start index of the answer (in the tokenized version of the input).
  * @property {number} [end] The character end index of the answer (in the tokenized version of the input).
  * @property {string} answer The answer to the question.
- * 
+ *
  * @typedef {Object} QuestionAnsweringPipelineOptions Parameters specific to question answering pipelines.
  * @property {number} [top_k=1] The number of top answer predictions to be returned.
- * 
+ *
  * @callback QuestionAnsweringPipelineCallback Answer the question(s) given as inputs by using the context(s).
  * @param {string|string[]} question One or several question(s) (must be used in conjunction with the `context` argument).
  * @param {string|string[]} context One or several context(s) associated with the question(s) (must be used in conjunction with the `question` argument).
  * @param {QuestionAnsweringPipelineOptions} [options] The options to use for question answering.
  * @returns {Promise<QuestionAnsweringOutput|QuestionAnsweringOutput[]>} An array or object containing the predicted answers and scores.
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & QuestionAnsweringPipelineCallback & Disposable} QuestionAnsweringPipelineType
  */
 
 /**
  * Question Answering pipeline using any `ModelForQuestionAnswering`.
- * 
+ *
  * **Example:** Run question answering with `Xenova/distilbert-base-uncased-distilled-squad`.
  * ```javascript
  * const answerer = await pipeline('question-answering', 'Xenova/distilbert-base-uncased-distilled-squad');
@@ -20548,10 +20714,10 @@ class QuestionAnsweringPipeline extends (/** @type {new (options: TextPipelineCo
  * @property {number} token The predicted token id (to replace the masked one).
  * @property {string} token_str The predicted token (to replace the masked one).
  * @typedef {FillMaskSingle[]} FillMaskOutput
- * 
+ *
  * @typedef {Object} FillMaskPipelineOptions Parameters specific to fill mask pipelines.
  * @property {number} [top_k=5] When passed, overrides the number of predictions to return.
- * 
+ *
  * @callback FillMaskPipelineCallback Fill the masked token in the text(s) given as inputs.
  * @param {string|string[]} texts One or several texts (or one list of prompts) with masked tokens.
  * @param {FillMaskPipelineOptions} [options] The options to use for masked language modelling.
@@ -20559,13 +20725,13 @@ class QuestionAnsweringPipeline extends (/** @type {new (options: TextPipelineCo
  * and the sequence with the predicted token filled in, or an array of such arrays (one for each input text).
  * If only one input text is given, the output will be an array of objects.
  * @throws {Error} When the mask token is not found in the input text.
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & FillMaskPipelineCallback & Disposable} FillMaskPipelineType
  */
 
 /**
  * Masked language modeling prediction pipeline using any `ModelWithLMHead`.
- * 
+ *
  * **Example:** Perform masked language modelling (a.k.a. "fill-mask") with `Xenova/bert-base-uncased`.
  * ```javascript
  * const unmasker = await pipeline('fill-mask', 'Xenova/bert-base-cased');
@@ -20578,7 +20744,7 @@ class QuestionAnsweringPipeline extends (/** @type {new (options: TextPipelineCo
  * //   { token_str: 'life', score: 0.01859794743359089, token: 1297, sequence: 'The goal of life is life.' }
  * // ]
  * ```
- * 
+ *
  * **Example:** Perform masked language modelling (a.k.a. "fill-mask") with `Xenova/bert-base-cased` (and return top result).
  * ```javascript
  * const unmasker = await pipeline('fill-mask', 'Xenova/bert-base-cased');
@@ -20655,18 +20821,18 @@ class FillMaskPipeline extends (/** @type {new (options: TextPipelineConstructor
  * @typedef {Object} Text2TextGenerationSingle
  * @property {string} generated_text The generated text.
  * @typedef {Text2TextGenerationSingle[]} Text2TextGenerationOutput
- * 
+ *
  * @callback Text2TextGenerationPipelineCallback Generate the output text(s) using text(s) given as inputs.
  * @param {string|string[]} texts Input text for the encoder.
  * @param {Partial<import('./generation/configuration_utils.js').GenerationConfig>} [options] Additional keyword arguments to pass along to the generate method of the model.
  * @returns {Promise<Text2TextGenerationOutput|Text2TextGenerationOutput[]>}
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & Text2TextGenerationPipelineCallback & Disposable} Text2TextGenerationPipelineType
  */
 
 /**
  * Text2TextGenerationPipeline class for generating text using a model that performs text-to-text generation tasks.
- * 
+ *
  * **Example:** Text-to-text generation w/ `Xenova/LaMini-Flan-T5-783M`.
  * ```javascript
  * const generator = await pipeline('text2text-generation', 'Xenova/LaMini-Flan-T5-783M');
@@ -20742,18 +20908,18 @@ class Text2TextGenerationPipeline extends (/** @type {new (options: TextPipeline
  * @typedef {Object} SummarizationSingle
  * @property {string} summary_text The summary text.
  * @typedef {SummarizationSingle[]} SummarizationOutput
- * 
+ *
  * @callback SummarizationPipelineCallback Summarize the text(s) given as inputs.
  * @param {string|string[]} texts One or several articles (or one list of articles) to summarize.
  * @param {import('./generation/configuration_utils.js').GenerationConfig} [options] Additional keyword arguments to pass along to the generate method of the model.
  * @returns {Promise<SummarizationOutput|SummarizationOutput[]>}
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & SummarizationPipelineCallback & Disposable} SummarizationPipelineType
  */
 
 /**
  * A pipeline for summarization tasks, inheriting from Text2TextGenerationPipeline.
- * 
+ *
  * **Example:** Summarization w/ `Xenova/distilbart-cnn-6-6`.
  * ```javascript
  * const generator = await pipeline('summarization', 'Xenova/distilbart-cnn-6-6');
@@ -20789,23 +20955,23 @@ class SummarizationPipeline extends (/** @type {new (options: TextPipelineConstr
  * @typedef {Object} TranslationSingle
  * @property {string} translation_text The translated text.
  * @typedef {TranslationSingle[]} TranslationOutput
- * 
+ *
  * @callback TranslationPipelineCallback Translate the text(s) given as inputs.
  * @param {string|string[]} texts Texts to be translated.
  * @param {import('./generation/configuration_utils.js').GenerationConfig} [options] Additional keyword arguments to pass along to the generate method of the model.
  * @returns {Promise<TranslationOutput|TranslationOutput[]>}
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & TranslationPipelineCallback & Disposable} TranslationPipelineType
  */
 
 /**
  * Translates text from one language to another.
- * 
+ *
  * **Example:** Multilingual translation w/ `Xenova/nllb-200-distilled-600M`.
- * 
+ *
  * See [here](https://github.com/facebookresearch/flores/blob/main/flores200/README.md#languages-in-flores-200)
  * for the full list of languages and their corresponding codes.
- * 
+ *
  * ```javascript
  * const translator = await pipeline('translation', 'Xenova/nllb-200-distilled-600M');
  * const output = await translator('जीवन एक चॉकलेट बॉक्स की तरह है।', {
@@ -20814,12 +20980,12 @@ class SummarizationPipeline extends (/** @type {new (options: TextPipelineConstr
  * });
  * // [{ translation_text: 'La vie est comme une boîte à chocolat.' }]
  * ```
- * 
+ *
  * **Example:** Multilingual translation w/ `Xenova/m2m100_418M`.
- * 
+ *
  * See [here](https://huggingface.co/facebook/m2m100_418M#languages-covered)
  * for the full list of languages and their corresponding codes.
- * 
+ *
  * ```javascript
  * const translator = await pipeline('translation', 'Xenova/m2m100_418M');
  * const output = await translator('生活就像一盒巧克力。', {
@@ -20828,12 +20994,12 @@ class SummarizationPipeline extends (/** @type {new (options: TextPipelineConstr
  * });
  * // [{ translation_text: 'Life is like a box of chocolate.' }]
  * ```
- * 
+ *
  * **Example:** Multilingual translation w/ `Xenova/mbart-large-50-many-to-many-mmt`.
- * 
+ *
  * See [here](https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt#languages-covered)
  * for the full list of languages and their corresponding codes.
- * 
+ *
  * ```javascript
  * const translator = await pipeline('translation', 'Xenova/mbart-large-50-many-to-many-mmt');
  * const output = await translator('संयुक्त राष्ट्र के प्रमुख का कहना है कि सीरिया में कोई सैन्य समाधान नहीं है', {
@@ -20862,21 +21028,21 @@ function isChat(x) {
 
 /**
  * @typedef {import('./tokenizers.js').Message[]} Chat
- * 
+ *
  * @typedef {Object} TextGenerationSingle
  * @property {string|Chat} generated_text The generated text.
  * @typedef {TextGenerationSingle[]} TextGenerationOutput
- * 
+ *
  * @typedef {Object} TextGenerationSpecificParams Parameters specific to text-generation pipelines.
  * @property {boolean} [add_special_tokens] Whether or not to add special tokens when tokenizing the sequences.
  * @property {boolean} [return_full_text=true] If set to `false` only added text is returned, otherwise the full text is returned.
  * @typedef {import('./generation/configuration_utils.js').GenerationConfig & TextGenerationSpecificParams} TextGenerationConfig
- * 
+ *
  * @callback TextGenerationPipelineCallback Complete the prompt(s) given as inputs.
  * @param {string|string[]|Chat|Chat[]} texts One or several prompts (or one list of prompts) to complete.
  * @param {Partial<TextGenerationConfig>} [options] Additional keyword arguments to pass along to the generate method of the model.
  * @returns {Promise<TextGenerationOutput|TextGenerationOutput[]>} An array or object containing the generated texts.
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & TextGenerationPipelineCallback & Disposable} TextGenerationPipelineType
  */
 
@@ -20884,7 +21050,7 @@ function isChat(x) {
  * Language generation pipeline using any `ModelWithLMHead` or `ModelForCausalLM`.
  * This pipeline predicts the words that will follow a specified text prompt.
  * NOTE: For the full list of generation parameters, see [`GenerationConfig`](./utils/generation#module_utils/generation.GenerationConfig).
- * 
+ *
  * **Example:** Text generation with `Xenova/distilgpt2` (default settings).
  * ```javascript
  * const generator = await pipeline('text-generation', 'Xenova/distilgpt2');
@@ -20892,7 +21058,7 @@ function isChat(x) {
  * const output = await generator(text);
  * // [{ generated_text: "I enjoy walking with my cute dog, and I love to play with the other dogs." }]
  * ```
- * 
+ *
  * **Example:** Text generation with `Xenova/distilgpt2` (custom settings).
  * ```javascript
  * const generator = await pipeline('text-generation', 'Xenova/distilgpt2');
@@ -20911,7 +21077,7 @@ function isChat(x) {
  * //   "generated_text": "Once upon a time, there was an abundance of information about the most important and influential"
  * // }]
  * ```
- * 
+ *
  * **Example:** Run code generation with `Xenova/codegen-350M-mono`.
  * ```javascript
  * const generator = await pipeline('text-generation', 'Xenova/codegen-350M-mono');
@@ -21030,7 +21196,7 @@ class TextGenerationPipeline extends (/** @type {new (options: TextPipelineConst
  * @property {string} sequence The sequence for which this is the output.
  * @property {string[]} labels The labels sorted by order of likelihood.
  * @property {number[]} scores The probabilities for each of the labels.
- * 
+ *
  * @typedef {Object} ZeroShotClassificationPipelineOptions Parameters specific to zero-shot classification pipelines.
  * @property {string} [hypothesis_template="This example is {}."] The template used to turn each
  * candidate label into an NLI-style hypothesis. The candidate label will replace the {} placeholder.
@@ -21038,14 +21204,14 @@ class TextGenerationPipeline extends (/** @type {new (options: TextPipelineConst
  * If `false`, the scores are normalized such that the sum of the label likelihoods for each sequence
  * is 1. If `true`, the labels are considered independent and probabilities are normalized for each
  * candidate by doing a softmax of the entailment score vs. the contradiction score.
- * 
+ *
  * @callback ZeroShotClassificationPipelineCallback Classify the sequence(s) given as inputs.
  * @param {string|string[]} texts The sequence(s) to classify, will be truncated if the model input is too large.
  * @param {string|string[]} candidate_labels The set of possible class labels to classify each sequence into.
  * Can be a single label, a string of comma-separated labels, or a list of labels.
  * @param {ZeroShotClassificationPipelineOptions} [options] The options to use for zero-shot classification.
  * @returns {Promise<ZeroShotClassificationOutput|ZeroShotClassificationOutput[]>} An array or object containing the predicted labels and scores.
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & ZeroShotClassificationPipelineCallback & Disposable} ZeroShotClassificationPipelineType
  */
 
@@ -21054,7 +21220,7 @@ class TextGenerationPipeline extends (/** @type {new (options: TextPipelineConst
  * trained on NLI (natural language inference) tasks. Equivalent of `text-classification`
  * pipelines, but these models don't require a hardcoded number of potential classes, they
  * can be chosen at runtime. It usually means it's slower but it is **much** more flexible.
- * 
+ *
  * **Example:** Zero shot classification with `Xenova/mobilebert-uncased-mnli`.
  * ```javascript
  * const classifier = await pipeline('zero-shot-classification', 'Xenova/mobilebert-uncased-mnli');
@@ -21067,7 +21233,7 @@ class TextGenerationPipeline extends (/** @type {new (options: TextPipelineConst
  * //   scores: [ 0.5562091040482018, 0.1843621307860853, 0.13942646639336376, 0.12000229877234923 ]
  * // }
  * ```
- * 
+ *
  * **Example:** Zero shot classification with `Xenova/nli-deberta-v3-xsmall` (multi-label).
  * ```javascript
  * const classifier = await pipeline('zero-shot-classification', 'Xenova/nli-deberta-v3-xsmall');
@@ -21181,20 +21347,20 @@ class ZeroShotClassificationPipeline extends (/** @type {new (options: TextPipel
  * @property {'none'|'mean'|'cls'} [pooling="none"] The pooling method to use.
  * @property {boolean} [normalize=false] Whether or not to normalize the embeddings in the last dimension.
  * @property {boolean} [quantize=false] Whether or not to quantize the embeddings.
- * @property {'binary'|'ubinary'} [precision='binary'] The precision to use for quantization. 
- * 
+ * @property {'binary'|'ubinary'} [precision='binary'] The precision to use for quantization.
+ *
  * @callback FeatureExtractionPipelineCallback Extract the features of the input(s).
  * @param {string|string[]} texts One or several texts (or one list of texts) to get the features of.
  * @param {FeatureExtractionPipelineOptions} [options] The options to use for feature extraction.
  * @returns {Promise<Tensor>} The features computed by the model.
- * 
+ *
  * @typedef {TextPipelineConstructorArgs & FeatureExtractionPipelineCallback & Disposable} FeatureExtractionPipelineType
  */
 
 /**
  * Feature extraction pipeline using no model head. This pipeline extracts the hidden
  * states from the base transformer, which can be used as features in downstream tasks.
- * 
+ *
  * **Example:** Run feature extraction with `bert-base-uncased` (without pooling/normalization).
  * ```javascript
  * const extractor = await pipeline('feature-extraction', 'Xenova/bert-base-uncased', { revision: 'default' });
@@ -21205,7 +21371,7 @@ class ZeroShotClassificationPipeline extends (/** @type {new (options: TextPipel
  * //   dims: [1, 8, 768]
  * // }
  * ```
- * 
+ *
  * **Example:** Run feature extraction with `bert-base-uncased` (with pooling/normalization).
  * ```javascript
  * const extractor = await pipeline('feature-extraction', 'Xenova/bert-base-uncased', { revision: 'default' });
@@ -21216,7 +21382,7 @@ class ZeroShotClassificationPipeline extends (/** @type {new (options: TextPipel
  * //   dims: [1, 768]
  * // }
  * ```
- * 
+ *
  * **Example:** Calculating embeddings with `sentence-transformers` models.
  * ```javascript
  * const extractor = await pipeline('feature-extraction', 'Xenova/all-MiniLM-L6-v2');
@@ -21297,19 +21463,19 @@ class FeatureExtractionPipeline extends (/** @type {new (options: TextPipelineCo
 /**
  * @typedef {Object} ImageFeatureExtractionPipelineOptions Parameters specific to image feature extraction pipelines.
  * @property {boolean} [pool=null] Whether or not to return the pooled output. If set to `false`, the model will return the raw hidden states.
- * 
+ *
  * @callback ImageFeatureExtractionPipelineCallback Extract the features of the input(s).
  * @param {ImagePipelineInputs} images One or several images (or one list of images) to get the features of.
  * @param {ImageFeatureExtractionPipelineOptions} [options] The options to use for image feature extraction.
  * @returns {Promise<Tensor>} The image features computed by the model.
- * 
+ *
  * @typedef {ImagePipelineConstructorArgs & ImageFeatureExtractionPipelineCallback & Disposable} ImageFeatureExtractionPipelineType
  */
 
 /**
  * Image feature extraction pipeline using no model head. This pipeline extracts the hidden
  * states from the base transformer, which can be used as features in downstream tasks.
- * 
+ *
  * **Example:** Perform image feature extraction with `Xenova/vit-base-patch16-224-in21k`.
  * ```javascript
  * const image_feature_extractor = await pipeline('image-feature-extraction', 'Xenova/vit-base-patch16-224-in21k');
@@ -21322,7 +21488,7 @@ class FeatureExtractionPipeline extends (/** @type {new (options: TextPipelineCo
  * //   size: 151296
  * // }
  * ```
- * 
+ *
  * **Example:** Compute image embeddings with `Xenova/clip-vit-base-patch32`.
  * ```javascript
  * const image_feature_extractor = await pipeline('image-feature-extraction', 'Xenova/clip-vit-base-patch32');
@@ -21378,12 +21544,12 @@ class ImageFeatureExtractionPipeline extends (/** @type {new (options: ImagePipe
  * @property {string} label The label predicted.
  * @property {number} score The corresponding probability.
  * @typedef {AudioClassificationSingle[]} AudioClassificationOutput
- * 
+ *
  * @typedef {Object} AudioClassificationPipelineOptions Parameters specific to audio classification pipelines.
  * @property {number} [top_k=5] The number of top labels that will be returned by the pipeline.
  * If the provided number is `null` or higher than the number of labels available in the model configuration,
  * it will default to the number of labels.
- * 
+ *
  * @callback AudioClassificationPipelineCallback Classify the sequence(s) given as inputs.
  * @param {AudioPipelineInputs} audio The input audio file(s) to be classified. The input is either:
  * - `string` or `URL` that is the filename/URL of the audio file, the file will be read at the processor's sampling rate
@@ -21392,14 +21558,14 @@ class ImageFeatureExtractionPipeline extends (/** @type {new (options: ImagePipe
  * - `Float32Array` or `Float64Array` of shape `(n, )`, representing the raw audio at the correct sampling rate (no further check will be done).
  * @param {AudioClassificationPipelineOptions} [options] The options to use for audio classification.
  * @returns {Promise<AudioClassificationOutput|AudioClassificationOutput[]>} An array or object containing the predicted labels and scores.
- * 
+ *
  * @typedef {AudioPipelineConstructorArgs & AudioClassificationPipelineCallback & Disposable} AudioClassificationPipelineType
  */
 
 /**
  * Audio classification pipeline using any `AutoModelForAudioClassification`.
  * This pipeline predicts the class of a raw waveform or an audio file.
- * 
+ *
  * **Example:** Perform audio classification with `Xenova/wav2vec2-large-xlsr-53-gender-recognition-librispeech`.
  * ```javascript
  * const classifier = await pipeline('audio-classification', 'Xenova/wav2vec2-large-xlsr-53-gender-recognition-librispeech');
@@ -21410,7 +21576,7 @@ class ImageFeatureExtractionPipeline extends (/** @type {new (options: ImagePipe
  * //   { label: 'female', score: 0.001845747814513743 }
  * // ]
  * ```
- * 
+ *
  * **Example:** Perform audio classification with `Xenova/ast-finetuned-audioset-10-10-0.4593` and return top 4 results.
  * ```javascript
  * const classifier = await pipeline('audio-classification', 'Xenova/ast-finetuned-audioset-10-10-0.4593');
@@ -21475,12 +21641,12 @@ class AudioClassificationPipeline extends (/** @type {new (options: AudioPipelin
  * @typedef {Object} ZeroShotAudioClassificationOutput
  * @property {string} label The label identified by the model. It is one of the suggested `candidate_label`.
  * @property {number} score The score attributed by the model for that label (between 0 and 1).
- * 
+ *
  * @typedef {Object} ZeroShotAudioClassificationPipelineOptions Parameters specific to zero-shot audio classification pipelines.
  * @property {string} [hypothesis_template="This is a sound of {}."] The sentence used in conjunction with `candidate_labels`
  * to attempt the audio classification by replacing the placeholder with the candidate_labels.
  * Then likelihood is estimated by using `logits_per_audio`.
- * 
+ *
  * @callback ZeroShotAudioClassificationPipelineCallback Classify the sequence(s) given as inputs.
  * @param {AudioPipelineInputs} audio The input audio file(s) to be classified. The input is either:
  * - `string` or `URL` that is the filename/URL of the audio file, the file will be read at the processor's sampling rate
@@ -21490,14 +21656,14 @@ class AudioClassificationPipeline extends (/** @type {new (options: AudioPipelin
  * @param {string[]} candidate_labels The candidate labels for this audio.
  * @param {ZeroShotAudioClassificationPipelineOptions} [options] The options to use for zero-shot audio classification.
  * @returns {Promise<ZeroShotAudioClassificationOutput[]|ZeroShotAudioClassificationOutput[][]>} An array of objects containing the predicted labels and scores.
- * 
+ *
  * @typedef {TextAudioPipelineConstructorArgs & ZeroShotAudioClassificationPipelineCallback & Disposable} ZeroShotAudioClassificationPipelineType
  */
 
 /**
  * Zero shot audio classification pipeline using `ClapModel`. This pipeline predicts the class of an audio when you
  * provide an audio and a set of `candidate_labels`.
- * 
+ *
  * **Example**: Perform zero-shot audio classification with `Xenova/clap-htsat-unfused`.
  * ```javascript
  * const classifier = await pipeline('zero-shot-audio-classification', 'Xenova/clap-htsat-unfused');
@@ -21530,7 +21696,7 @@ class ZeroShotAudioClassificationPipeline extends (/** @type {new (options: Text
             audio = [/** @type {AudioInput} */ (audio)];
         }
 
-        // Insert label into hypothesis template 
+        // Insert label into hypothesis template
         const texts = candidate_labels.map(
             x => hypothesis_template.replace('{}', x)
         );
@@ -21574,7 +21740,7 @@ class ZeroShotAudioClassificationPipeline extends (/** @type {new (options: Text
  * @property {string} text The recognized text.
  * @property {Chunk[]} [chunks] When using `return_timestamps`, the `chunks` will become a list
  * containing all the various text chunks identified by the model.
- * 
+ *
  * @typedef {Object} AutomaticSpeechRecognitionSpecificParams Parameters specific to automatic-speech-recognition pipelines.
  * @property {boolean|'word'} [return_timestamps] Whether to return timestamps or not. Default is `false`.
  * @property {number} [chunk_length_s] The length of audio chunks to process in seconds. Default is 0 (no chunking).
@@ -21584,7 +21750,7 @@ class ZeroShotAudioClassificationPipeline extends (/** @type {new (options: Text
  * @property {string} [task] The task to perform. Default is `null`, meaning it should be auto-detected.
  * @property {number} [num_frames] The number of frames in the input audio.
  * @typedef {import('./generation/configuration_utils.js').GenerationConfig & AutomaticSpeechRecognitionSpecificParams} AutomaticSpeechRecognitionConfig
- * 
+ *
  * @callback AutomaticSpeechRecognitionPipelineCallback Transcribe the audio sequence(s) given as inputs to text.
  * @param {AudioPipelineInputs} audio The input audio file(s) to be transcribed. The input is either:
  * - `string` or `URL` that is the filename/URL of the audio file, the file will be read at the processor's sampling rate
@@ -21593,7 +21759,7 @@ class ZeroShotAudioClassificationPipeline extends (/** @type {new (options: Text
  * - `Float32Array` or `Float64Array` of shape `(n, )`, representing the raw audio at the correct sampling rate (no further check will be done).
  * @param {Partial<AutomaticSpeechRecognitionConfig>} [options] Additional keyword arguments to pass along to the generate method of the model.
  * @returns {Promise<AutomaticSpeechRecognitionOutput|AutomaticSpeechRecognitionOutput[]>} An object containing the transcription text and optionally timestamps if `return_timestamps` is `true`.
- * 
+ *
  * @typedef {TextAudioPipelineConstructorArgs & AutomaticSpeechRecognitionPipelineCallback & Disposable} AutomaticSpeechRecognitionPipelineType
  */
 
@@ -21607,7 +21773,7 @@ class ZeroShotAudioClassificationPipeline extends (/** @type {new (options: Text
  * const output = await transcriber(url);
  * // { text: " And so my fellow Americans ask not what your country can do for you, ask what you can do for your country." }
  * ```
- * 
+ *
  * **Example:** Transcribe English w/ timestamps.
  * ```javascript
  * const transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-tiny.en');
@@ -21621,7 +21787,7 @@ class ZeroShotAudioClassificationPipeline extends (/** @type {new (options: Text
  * //   ]
  * // }
  * ```
- * 
+ *
  * **Example:** Transcribe English w/ word-level timestamps.
  * ```javascript
  * const transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-tiny.en');
@@ -21640,7 +21806,7 @@ class ZeroShotAudioClassificationPipeline extends (/** @type {new (options: Text
  * //   ]
  * // }
  * ```
- * 
+ *
  * **Example:** Transcribe French.
  * ```javascript
  * const transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-small');
@@ -21648,7 +21814,7 @@ class ZeroShotAudioClassificationPipeline extends (/** @type {new (options: Text
  * const output = await transcriber(url, { language: 'french', task: 'transcribe' });
  * // { text: " J'adore, j'aime, je n'aime pas, je déteste." }
  * ```
- * 
+ *
  * **Example:** Translate French to English.
  * ```javascript
  * const transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-small');
@@ -21656,7 +21822,7 @@ class ZeroShotAudioClassificationPipeline extends (/** @type {new (options: Text
  * const output = await transcriber(url, { language: 'french', task: 'translate' });
  * // { text: " I love, I like, I don't like, I hate." }
  * ```
- * 
+ *
  * **Example:** Transcribe/translate audio longer than 30 seconds.
  * ```javascript
  * const transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-tiny.en');
@@ -21879,18 +22045,18 @@ class AutomaticSpeechRecognitionPipeline extends (/** @type {new (options: TextA
  * @typedef {Object} ImageToTextSingle
  * @property {string} generated_text The generated text.
  * @typedef {ImageToTextSingle[]} ImageToTextOutput
- * 
+ *
  * @callback ImageToTextPipelineCallback Assign labels to the image(s) passed as inputs.
  * @param {ImagePipelineInputs} texts The images to be captioned.
  * @param {Partial<import('./generation/configuration_utils.js').GenerationConfig>} [options] Additional keyword arguments to pass along to the generate method of the model.
  * @returns {Promise<ImageToTextOutput|ImageToTextOutput[]>} An object (or array of objects) containing the generated text(s).
- * 
+ *
  * @typedef {TextImagePipelineConstructorArgs & ImageToTextPipelineCallback & Disposable} ImageToTextPipelineType
  */
 
 /**
  * Image To Text pipeline using a `AutoModelForVision2Seq`. This pipeline predicts a caption for a given image.
- * 
+ *
  * **Example:** Generate a caption for an image w/ `Xenova/vit-gpt2-image-captioning`.
  * ```javascript
  * const captioner = await pipeline('image-to-text', 'Xenova/vit-gpt2-image-captioning');
@@ -21898,7 +22064,7 @@ class AutomaticSpeechRecognitionPipeline extends (/** @type {new (options: TextA
  * const output = await captioner(url);
  * // [{ generated_text: 'a cat laying on a couch with another cat' }]
  * ```
- * 
+ *
  * **Example:** Optical Character Recognition (OCR) w/ `Xenova/trocr-small-handwritten`.
  * ```javascript
  * const captioner = await pipeline('image-to-text', 'Xenova/trocr-small-handwritten');
@@ -21944,22 +22110,22 @@ class ImageToTextPipeline extends (/** @type {new (options: TextImagePipelineCon
  * @property {string} label The label identified by the model.
  * @property {number} score The score attributed by the model for that label.
  * @typedef {ImageClassificationSingle[]} ImageClassificationOutput
- * 
+ *
  * @typedef {Object} ImageClassificationPipelineOptions Parameters specific to image classification pipelines.
- * @property {number} [top_k=1] The number of top labels that will be returned by the pipeline. 
- * 
+ * @property {number} [top_k=1] The number of top labels that will be returned by the pipeline.
+ *
  * @callback ImageClassificationPipelineCallback Assign labels to the image(s) passed as inputs.
  * @param {ImagePipelineInputs} images The input images(s) to be classified.
  * @param {ImageClassificationPipelineOptions} [options] The options to use for image classification.
  * @returns {Promise<ImageClassificationOutput|ImageClassificationOutput[]>} An array or object containing the predicted labels and scores.
- * 
+ *
  * @typedef {ImagePipelineConstructorArgs & ImageClassificationPipelineCallback & Disposable} ImageClassificationPipelineType
  */
 
 /**
  * Image classification pipeline using any `AutoModelForImageClassification`.
  * This pipeline predicts the class of an image.
- * 
+ *
  * **Example:** Classify an image.
  * ```javascript
  * const classifier = await pipeline('image-classification', 'Xenova/vit-base-patch16-224');
@@ -21969,7 +22135,7 @@ class ImageToTextPipeline extends (/** @type {new (options: TextImagePipelineCon
  * //   { label: 'tiger, Panthera tigris', score: 0.632695734500885 },
  * // ]
  * ```
- * 
+ *
  * **Example:** Classify an image and return top `n` classes.
  * ```javascript
  * const classifier = await pipeline('image-classification', 'Xenova/vit-base-patch16-224');
@@ -21981,7 +22147,7 @@ class ImageToTextPipeline extends (/** @type {new (options: TextImagePipelineCon
  * //   { label: 'lion, king of beasts, Panthera leo', score: 0.00045060308184474707 },
  * // ]
  * ```
- * 
+ *
  * **Example:** Classify an image and return all classes.
  * ```javascript
  * const classifier = await pipeline('image-classification', 'Xenova/vit-base-patch16-224');
@@ -22048,7 +22214,7 @@ class ImageClassificationPipeline extends (/** @type {new (options: ImagePipelin
  * @property {string|null} label The label of the segment.
  * @property {number|null} score The score of the segment.
  * @property {RawImage} mask The mask of the segment.
- * 
+ *
  * @typedef {Object} ImageSegmentationPipelineOptions Parameters specific to image segmentation pipelines.
  * @property {number} [threshold=0.5] Probability threshold to filter out predicted masks.
  * @property {number} [mask_threshold=0.5] Threshold to use when turning the predicted masks into binary values.
@@ -22057,19 +22223,19 @@ class ImageClassificationPipeline extends (/** @type {new (options: ImagePipelin
  * depending on model capabilities. If not set, the pipeline will attempt to resolve (in that order).
  * @property {number[]} [label_ids_to_fuse=null] List of label ids to fuse. If not set, do not fuse any labels.
  * @property {number[][]} [target_sizes=null] List of target sizes for the input images. If not set, use the original image sizes.
- * 
+ *
  * @callback ImageSegmentationPipelineCallback Segment the input images.
  * @param {ImagePipelineInputs} images The input images.
  * @param {ImageSegmentationPipelineOptions} [options] The options to use for image segmentation.
  * @returns {Promise<ImageSegmentationPipelineOutput[]>} The annotated segments.
- * 
+ *
  * @typedef {ImagePipelineConstructorArgs & ImageSegmentationPipelineCallback & Disposable} ImageSegmentationPipelineType
  */
 
 /**
  * Image segmentation pipeline using any `AutoModelForXXXSegmentation`.
  * This pipeline predicts masks of objects and their classes.
- * 
+ *
  * **Example:** Perform image segmentation with `Xenova/detr-resnet-50-panoptic`.
  * ```javascript
  * const segmenter = await pipeline('image-segmentation', 'Xenova/detr-resnet-50-panoptic');
@@ -22153,12 +22319,17 @@ class ImageSegmentationPipeline extends (/** @type {new (options: ImagePipelineC
         /** @type {ImageSegmentationPipelineOutput[]} */
         const annotation = [];
         if (!subtask) {
+            // We define an epsilon to safeguard against numerical/precision issues when detecting
+            // the normalization mode of the output (i.e., sigmoid already applied, or not).
+            // See https://github.com/microsoft/onnxruntime/issues/23943 for more information.
+            const epsilon = 1e-5;
+
             // Perform standard image segmentation
             const result = output[outputNames[0]];
             for (let i = 0; i < imageSizes.length; ++i) {
                 const size = imageSizes[i];
                 const item = result[i];
-                if (item.data.some(x => x < 0 || x > 1)) {
+                if (item.data.some(x => x < -epsilon || x > 1 + epsilon)) {
                     item.sigmoid_();
                 }
                 const mask = await _utils_image_js__WEBPACK_IMPORTED_MODULE_9__.RawImage.fromTensor(item.mul_(255).to('uint8')).resize(size[1], size[0]);
@@ -22227,19 +22398,19 @@ class ImageSegmentationPipeline extends (/** @type {new (options: ImagePipelineC
 
 /**
  * @typedef {Object} BackgroundRemovalPipelineOptions Parameters specific to image segmentation pipelines.
- * 
+ *
  * @callback BackgroundRemovalPipelineCallback Segment the input images.
  * @param {ImagePipelineInputs} images The input images.
  * @param {BackgroundRemovalPipelineOptions} [options] The options to use for image segmentation.
  * @returns {Promise<RawImage[]>} The images with the background removed.
- * 
+ *
  * @typedef {ImagePipelineConstructorArgs & BackgroundRemovalPipelineCallback & Disposable} BackgroundRemovalPipelineType
  */
 
 /**
  * Background removal pipeline using certain `AutoModelForXXXSegmentation`.
  * This pipeline removes the backgrounds of images.
- * 
+ *
  * **Example:** Perform background removal with `Xenova/modnet`.
  * ```javascript
  * const segmenter = await pipeline('background-removal', 'Xenova/modnet');
@@ -22250,7 +22421,7 @@ class ImageSegmentationPipeline extends (/** @type {new (options: ImagePipelineC
  * // ]
  * ```
  */
-class BackgroundRemovalPipeline extends (/** @type {new (options: ImagePipelineConstructorArgs) => ImageSegmentationPipelineType} */ (ImageSegmentationPipeline)) {
+class BackgroundRemovalPipeline extends (/** @type {new (options: ImagePipelineConstructorArgs) => BackgroundRemovalPipelineType} */ (/** @type {any} */(ImageSegmentationPipeline))) {
     /**
      * Create a new BackgroundRemovalPipeline.
      * @param {ImagePipelineConstructorArgs} options An object used to instantiate the pipeline.
@@ -22285,25 +22456,25 @@ class BackgroundRemovalPipeline extends (/** @type {new (options: ImagePipelineC
  * @typedef {Object} ZeroShotImageClassificationOutput
  * @property {string} label The label identified by the model. It is one of the suggested `candidate_label`.
  * @property {number} score The score attributed by the model for that label (between 0 and 1).
- * 
+ *
  * @typedef {Object} ZeroShotImageClassificationPipelineOptions Parameters specific to zero-shot image classification pipelines.
  * @property {string} [hypothesis_template="This is a photo of {}"] The sentence used in conjunction with `candidate_labels`
  * to attempt the image classification by replacing the placeholder with the candidate_labels.
  * Then likelihood is estimated by using `logits_per_image`.
- * 
+ *
  * @callback ZeroShotImageClassificationPipelineCallback Assign labels to the image(s) passed as inputs.
  * @param {ImagePipelineInputs} images The input images.
  * @param {string[]} candidate_labels The candidate labels for this image.
  * @param {ZeroShotImageClassificationPipelineOptions} [options] The options to use for zero-shot image classification.
  * @returns {Promise<ZeroShotImageClassificationOutput[]|ZeroShotImageClassificationOutput[][]>} An array of objects containing the predicted labels and scores.
- * 
+ *
  * @typedef {TextImagePipelineConstructorArgs & ZeroShotImageClassificationPipelineCallback & Disposable} ZeroShotImageClassificationPipelineType
  */
 
 /**
  * Zero shot image classification pipeline. This pipeline predicts the class of
  * an image when you provide an image and a set of `candidate_labels`.
- * 
+ *
  * **Example:** Zero shot image classification w/ `Xenova/clip-vit-base-patch32`.
  * ```javascript
  * const classifier = await pipeline('zero-shot-image-classification', 'Xenova/clip-vit-base-patch32');
@@ -22333,7 +22504,7 @@ class ZeroShotImageClassificationPipeline extends (/** @type {new (options: Text
         const isBatched = Array.isArray(images);
         const preparedImages = await prepareImages(images);
 
-        // Insert label into hypothesis template 
+        // Insert label into hypothesis template
         const texts = candidate_labels.map(
             x => hypothesis_template.replace('{}', x)
         );
@@ -22380,23 +22551,23 @@ class ZeroShotImageClassificationPipeline extends (/** @type {new (options: Text
  * @property {number} score The score attributed by the model for that label.
  * @property {BoundingBox} box The bounding box of detected object in image's original size, or as a percentage if `percentage` is set to true.
  * @typedef {ObjectDetectionPipelineSingle[]} ObjectDetectionPipelineOutput
- * 
+ *
  * @typedef {Object} ObjectDetectionPipelineOptions Parameters specific to object detection pipelines.
  * @property {number} [threshold=0.9] The threshold used to filter boxes by score.
  * @property {boolean} [percentage=false] Whether to return the boxes coordinates in percentage (true) or in pixels (false).
- * 
+ *
  * @callback ObjectDetectionPipelineCallback Detect objects (bounding boxes & classes) in the image(s) passed as inputs.
  * @param {ImagePipelineInputs} images The input images.
  * @param {ObjectDetectionPipelineOptions} [options] The options to use for object detection.
- * @returns {Promise<ObjectDetectionPipelineOutput|ObjectDetectionPipelineOutput[]>} A list of objects or a list of list of objects. 
- * 
+ * @returns {Promise<ObjectDetectionPipelineOutput|ObjectDetectionPipelineOutput[]>} A list of objects or a list of list of objects.
+ *
  * @typedef {ImagePipelineConstructorArgs & ObjectDetectionPipelineCallback & Disposable} ObjectDetectionPipelineType
  */
 
 /**
  * Object detection pipeline using any `AutoModelForObjectDetection`.
  * This pipeline predicts bounding boxes of objects and their classes.
- * 
+ *
  * **Example:** Run object-detection with `Xenova/detr-resnet-50`.
  * ```javascript
  * const detector = await pipeline('object-detection', 'Xenova/detr-resnet-50');
@@ -22470,27 +22641,27 @@ class ObjectDetectionPipeline extends (/** @type {new (options: ImagePipelineCon
  * @property {string} label Text query corresponding to the found object.
  * @property {number} score Score corresponding to the object (between 0 and 1).
  * @property {BoundingBox} box Bounding box of the detected object in image's original size, or as a percentage if `percentage` is set to true.
- * 
+ *
  * @typedef {Object} ZeroShotObjectDetectionPipelineOptions Parameters specific to zero-shot object detection pipelines.
  * @property {number} [threshold=0.1] The probability necessary to make a prediction.
  * @property {number} [top_k=null] The number of top predictions that will be returned by the pipeline.
  * If the provided number is `null` or higher than the number of predictions available, it will default
  * to the number of predictions.
  * @property {boolean} [percentage=false] Whether to return the boxes coordinates in percentage (true) or in pixels (false).
- * 
+ *
  * @callback ZeroShotObjectDetectionPipelineCallback Detect objects (bounding boxes & classes) in the image(s) passed as inputs.
  * @param {ImagePipelineInputs} images The input images.
  * @param {string[]} candidate_labels What the model should recognize in the image.
  * @param {ZeroShotObjectDetectionPipelineOptions} [options] The options to use for zero-shot object detection.
  * @returns {Promise<ZeroShotObjectDetectionOutput[]|ZeroShotObjectDetectionOutput[][]>} An array of objects containing the predicted labels, scores, and bounding boxes.
- * 
+ *
  * @typedef {TextImagePipelineConstructorArgs & ZeroShotObjectDetectionPipelineCallback & Disposable} ZeroShotObjectDetectionPipelineType
  */
 
 /**
  * Zero-shot object detection pipeline. This pipeline predicts bounding boxes of
  * objects when you provide an image and a set of `candidate_labels`.
- * 
+ *
  * **Example:** Zero-shot object detection w/ `Xenova/owlvit-base-patch32`.
  * ```javascript
  * const detector = await pipeline('zero-shot-object-detection', 'Xenova/owlvit-base-patch32');
@@ -22520,7 +22691,7 @@ class ObjectDetectionPipeline extends (/** @type {new (options: ImagePipelineCon
  * //   }
  * // ]
  * ```
- * 
+ *
  * **Example:** Zero-shot object detection w/ `Xenova/owlvit-base-patch32` (returning top 4 matches and setting a threshold).
  * ```javascript
  * const detector = await pipeline('zero-shot-object-detection', 'Xenova/owlvit-base-patch32');
@@ -22635,13 +22806,13 @@ class ZeroShotObjectDetectionPipeline extends (/** @type {new (options: TextImag
  * @typedef {Object} DocumentQuestionAnsweringSingle
  * @property {string} answer The generated text.
  * @typedef {DocumentQuestionAnsweringSingle[]} DocumentQuestionAnsweringOutput
- * 
+ *
  * @callback DocumentQuestionAnsweringPipelineCallback Answer the question given as input by using the document.
  * @param {ImageInput} image The image of the document to use.
  * @param {string} question A question to ask of the document.
  * @param {Partial<import('./generation/configuration_utils.js').GenerationConfig>} [options] Additional keyword arguments to pass along to the generate method of the model.
  * @returns {Promise<DocumentQuestionAnsweringOutput|DocumentQuestionAnsweringOutput[]>} An object (or array of objects) containing the answer(s).
- * 
+ *
  * @typedef {TextImagePipelineConstructorArgs & DocumentQuestionAnsweringPipelineCallback & Disposable} DocumentQuestionAnsweringPipelineType
  */
 
@@ -22649,7 +22820,7 @@ class ZeroShotObjectDetectionPipeline extends (/** @type {new (options: TextImag
  * Document Question Answering pipeline using any `AutoModelForDocumentQuestionAnswering`.
  * The inputs/outputs are similar to the (extractive) question answering pipeline; however,
  * the pipeline takes an image (and optional OCR'd words/boxes) as input instead of text context.
- * 
+ *
  * **Example:** Answer questions about a document with `Xenova/donut-base-finetuned-docvqa`.
  * ```javascript
  * const qa_pipeline = await pipeline('document-question-answering', 'Xenova/donut-base-finetuned-docvqa');
@@ -22719,22 +22890,22 @@ class DocumentQuestionAnsweringPipeline extends (/** @type {new (options: TextIm
  * @typedef {Object} TextToAudioOutput
  * @property {Float32Array} audio The generated audio waveform.
  * @property {number} sampling_rate The sampling rate of the generated audio waveform.
- * 
+ *
  * @typedef {Object} TextToAudioPipelineOptions Parameters specific to text-to-audio pipelines.
  * @property {Tensor|Float32Array|string|URL} [speaker_embeddings=null] The speaker embeddings (if the model requires it).
- * 
+ *
  * @callback TextToAudioPipelineCallback Generates speech/audio from the inputs.
  * @param {string|string[]} texts The text(s) to generate.
  * @param {TextToAudioPipelineOptions} options Parameters passed to the model generation/forward method.
  * @returns {Promise<TextToAudioOutput>} An object containing the generated audio and sampling rate.
- * 
+ *
  * @typedef {TextToAudioPipelineConstructorArgs & TextToAudioPipelineCallback & Disposable} TextToAudioPipelineType
  */
 
 /**
  * Text-to-audio generation pipeline using any `AutoModelForTextToWaveform` or `AutoModelForTextToSpectrogram`.
  * This pipeline generates an audio file from an input text and optional other conditional inputs.
- * 
+ *
  * **Example:** Generate audio from text with `Xenova/speecht5_tts`.
  * ```javascript
  * const synthesizer = await pipeline('text-to-speech', 'Xenova/speecht5_tts', { quantized: false });
@@ -22745,17 +22916,17 @@ class DocumentQuestionAnsweringPipeline extends (/** @type {new (options: TextIm
  * //   sampling_rate: 16000
  * // }
  * ```
- * 
+ *
  * You can then save the audio to a .wav file with the `wavefile` package:
  * ```javascript
  * import wavefile from 'wavefile';
  * import fs from 'fs';
- * 
+ *
  * const wav = new wavefile.WaveFile();
  * wav.fromScratch(1, out.sampling_rate, '32f', out.audio);
  * fs.writeFileSync('out.wav', wav.toBuffer());
  * ```
- * 
+ *
  * **Example:** Multilingual speech generation with `Xenova/mms-tts-fra`. See [here](https://huggingface.co/models?pipeline_tag=text-to-speech&other=vits&sort=trending) for the full list of available languages (1107).
  * ```javascript
  * const synthesizer = await pipeline('text-to-speech', 'Xenova/mms-tts-fra');
@@ -22861,13 +23032,13 @@ class TextToAudioPipeline extends (/** @type {new (options: TextToAudioPipelineC
  * @callback ImageToImagePipelineCallback Transform the image(s) passed as inputs.
  * @param {ImagePipelineInputs} images The images to transform.
  * @returns {Promise<RawImage|RawImage[]>} The transformed image or list of images.
- * 
+ *
  * @typedef {ImagePipelineConstructorArgs & ImageToImagePipelineCallback & Disposable} ImageToImagePipelineType
  */
 
 /**
  * Image to Image pipeline using any `AutoModelForImageToImage`. This pipeline generates an image based on a previous image input.
- * 
+ *
  * **Example:** Super-resolution w/ `Xenova/swin2SR-classical-sr-x2-64`
  * ```javascript
  * const upscaler = await pipeline('image-to-image', 'Xenova/swin2SR-classical-sr-x2-64');
@@ -22912,17 +23083,17 @@ class ImageToImagePipeline extends (/** @type {new (options: ImagePipelineConstr
  * @typedef {Object} DepthEstimationPipelineOutput
  * @property {Tensor} predicted_depth The raw depth map predicted by the model.
  * @property {RawImage} depth The processed depth map as an image (with the same size as the input image).
- * 
+ *
  * @callback DepthEstimationPipelineCallback Predicts the depth for the image(s) passed as inputs.
  * @param {ImagePipelineInputs} images The images to compute depth for.
  * @returns {Promise<DepthEstimationPipelineOutput|DepthEstimationPipelineOutput[]>} An image or a list of images containing result(s).
- * 
+ *
  * @typedef {ImagePipelineConstructorArgs & DepthEstimationPipelineCallback & Disposable} DepthEstimationPipelineType
  */
 
 /**
  * Depth estimation pipeline using any `AutoModelForDepthEstimation`. This pipeline predicts the depth of an image.
- * 
+ *
  * **Example:** Depth estimation w/ `Xenova/dpt-hybrid-midas`
  * ```javascript
  * const depth_estimator = await pipeline('depth-estimation', 'Xenova/dpt-hybrid-midas');
@@ -23307,7 +23478,7 @@ const TASK_ALIASES = Object.freeze({
 
 /**
  * Utility factory method to build a `Pipeline` object.
- * 
+ *
  * @template {PipelineType} T The type of pipeline to return.
  * @param {T} task The task defining which pipeline will be returned. Currently accepted tasks are:
  *  - `"audio-classification"`: will return a `AudioClassificationPipeline`.
@@ -26126,13 +26297,12 @@ class PreTrainedTokenizer extends _utils_generic_js__WEBPACK_IMPORTED_MODULE_0__
             this.decoder.end_of_word_suffix = this.model.end_of_word_suffix;
         }
 
-        this.added_tokens_regex = this.added_tokens.length > 0 ? new RegExp(
-            this.added_tokens.slice()
-                // Sort by length (desc) to avoid early partial matches
-                .sort((a, b) => b.content.length - a.content.length)
-                .map(x => `${x.lstrip ? '\\s*' : ''}(${(0,_utils_core_js__WEBPACK_IMPORTED_MODULE_1__.escapeRegExp)(x.content)})${x.rstrip ? '\\s*' : ''}`)
-                .join('|')
-        ) : null;
+        this.added_tokens_splitter = new _utils_data_structures_js__WEBPACK_IMPORTED_MODULE_5__.DictionarySplitter(
+            this.added_tokens.map(x => x.content),
+        );
+
+        /** @type {Map<string, AddedToken>} */
+        this.added_tokens_map = new Map(this.added_tokens.map(x => [x.content, x]))
 
         // Set mask token if present (otherwise will be undefined, which is fine)
         this.mask_token = this.getToken('mask_token');
@@ -26427,40 +26597,50 @@ class PreTrainedTokenizer extends _utils_generic_js__WEBPACK_IMPORTED_MODULE_0__
         // Actual function which does encoding, for a single text
         // First, we take care of special tokens. Needed to avoid issues arising from
         // normalization and/or pretokenization (which may not preserve special tokens)
-        const sections = this.added_tokens_regex ? text.split(this.added_tokens_regex).filter(x => x) : [text];
-
-        const tokens = sections.map((x, section_index) => {
-            const addedToken = this.added_tokens.find(t => t.content === x);
-            if (addedToken !== undefined) {
-                // Ignore added tokens
-                return x
-            } else {
-                if (this.remove_space === true) {
-                    x = x.trim().split(/\s+/).join(' ');
-                }
-                if (this.do_lowercase_and_remove_accent) {
-                    x = lowercase_and_remove_accent(x);
+        const sections = this.added_tokens_splitter.split(text);
+
+        // Process left/right stripping of added tokens
+        for (let i = 0; i < sections.length; ++i) {
+            const addedToken = this.added_tokens_map.get(sections[i]);
+            if (addedToken) {
+                if (addedToken.lstrip && i > 0) {
+                    sections[i - 1] = sections[i - 1].trimEnd();
                 }
-
-                if (this.normalizer !== null) {
-                    x = this.normalizer(x);
+                if (addedToken.rstrip && i < sections.length - 1) {
+                    sections[i + 1] = sections[i + 1].trimStart();
                 }
+            }
+        }
 
-                // If, after normalization, this section is empty (e.g., trimming whitespace),
-                // we return an empty array
-                if (x.length === 0) {
-                    return [];
-                }
+        const tokens = sections.flatMap((x, section_index) => {
+            if (x.length === 0) return [];
+            if (this.added_tokens_map.has(x)) return [x]; // Return added tokens unchanged
 
-                const sectionTokens = (this.pre_tokenizer !== null) ? this.pre_tokenizer(x, {
-                    section_index,
-                }) : [x];
+            if (this.remove_space === true) {
+                x = x.trim().split(/\s+/).join(' ');
+            }
+            if (this.do_lowercase_and_remove_accent) {
+                x = lowercase_and_remove_accent(x);
+            }
 
-                const tokens = this.model(sectionTokens);
+            if (this.normalizer !== null) {
+                x = this.normalizer(x);
+            }
 
-                return tokens;
+            // If, after normalization, this section is empty (e.g., trimming whitespace),
+            // we return an empty array
+            if (x.length === 0) {
+                return [];
             }
-        }).flat();
+
+            const sectionTokens = (this.pre_tokenizer !== null) ? this.pre_tokenizer(x, {
+                section_index,
+            }) : [x];
+
+            const tokens = this.model(sectionTokens);
+
+            return tokens;
+        });
 
         return tokens;
     }
@@ -29089,6 +29269,7 @@ function count(arr, value) {
 __webpack_require__.r(__webpack_exports__);
 /* harmony export */ __webpack_require__.d(__webpack_exports__, {
 /* harmony export */   CharTrie: () => (/* binding */ CharTrie),
+/* harmony export */   DictionarySplitter: () => (/* binding */ DictionarySplitter),
 /* harmony export */   PriorityQueue: () => (/* binding */ PriorityQueue),
 /* harmony export */   TokenLattice: () => (/* binding */ TokenLattice)
 /* harmony export */ });
@@ -29540,6 +29721,80 @@ class TokenLatticeNode {
     }
 }
 
+/**
+ * A data structure which uses a trie to split a string into tokens based on a dictionary.
+ * It can also use a regular expression to preprocess the input text before splitting.
+ * 
+ * NOTE: To ensure multi-byte characters are handled correctly, we operate at byte-level instead of character-level.
+ */
+class DictionarySplitter {
+    /**
+     * @param {string[]} dictionary The dictionary of words to use for splitting.
+     */
+    constructor(dictionary) {
+        this.trie = this._buildTrie(dictionary);
+    }
+
+    /**
+     * Builds a trie from the given dictionary.
+     * @param {string[]} dictionary The dictionary of words to build the trie from.
+     * @returns {Object} The root node of the trie.
+     * @private
+     */
+    _buildTrie(dictionary) {
+        const trie = Object.create(null);
+        for (const word of dictionary) {
+            let node = trie;
+            for (let i = 0; i < word.length; ++i) {
+                node = (node[word[i]] ??= Object.create(null));
+            }
+            node.end = word;
+        }
+        return trie;
+    }
+
+    /**
+     * Splits the input text into tokens based on the dictionary.
+     * @param {string} text The input text to split.
+     * @returns {string[]} An array of tokens.
+     */
+    split(text) {
+        const result = [];
+        const n = text.length;
+        let start = 0;
+        let i = 0;
+
+        while (i < n) {
+            let node = this.trie;
+            let match = null;
+            let j = i;
+
+            while (j < n && (node = node[text[j]])) {
+                if (node.end) {
+                    // Always keep the last (i.e., longest) match.
+                    match = node.end;
+                }
+                ++j;
+            }
+
+            if (match) {
+                if (i > start) {
+                    result.push(text.slice(start, i));
+                }
+                result.push(match);
+                i += match.length;
+                start = i;
+            } else {
+                ++i;
+            }
+        }
+        if (start < n) {
+            result.push(text.slice(start));
+        }
+        return result;
+    }
+}
+
 
 /***/ }),
 
@@ -29923,6 +30178,22 @@ function isValidUrl(string, protocols = null, validHosts = null) {
     return true;
 }
 
+const REPO_ID_REGEX = /^(\b[\w\-.]+\b\/)?\b[\w\-.]{1,96}\b$/;
+
+/**
+ * Tests whether a string is a valid Hugging Face model ID or not.
+ * Adapted from https://github.com/huggingface/huggingface_hub/blob/6378820ebb03f071988a96c7f3268f5bdf8f9449/src/huggingface_hub/utils/_validators.py#L119-L170
+ * 
+ * @param {string} string The string to test
+ * @returns {boolean} True if the string is a valid model ID, false otherwise.
+ */
+function isValidHfModelId(string) {
+    if (!REPO_ID_REGEX.test(string)) return false;
+    if (string.includes("..") || string.includes("--")) return false;
+    if (string.endsWith(".git") || string.endsWith(".ipynb")) return false;
+    return true;
+}
+
 /**
  * Helper function to get a file, using either the Fetch API or FileSystem API.
  *
@@ -30175,12 +30446,13 @@ async function getModelFile(path_or_repo_id, filename, fatal = true, options = {
     }
 
     const revision = options.revision ?? 'main';
+    const requestURL = pathJoin(path_or_repo_id, filename);
 
-    let requestURL = pathJoin(path_or_repo_id, filename);
-    let cachePath = pathJoin(_env_js__WEBPACK_IMPORTED_MODULE_2__.env.localModelPath, requestURL);
-
-    let localPath = requestURL;
-    let remoteURL = pathJoin(
+    const validModelId = isValidHfModelId(path_or_repo_id);
+    const localPath = validModelId
+        ? pathJoin(_env_js__WEBPACK_IMPORTED_MODULE_2__.env.localModelPath, requestURL)
+        : requestURL;
+    const remoteURL = pathJoin(
         _env_js__WEBPACK_IMPORTED_MODULE_2__.env.remoteHost,
         _env_js__WEBPACK_IMPORTED_MODULE_2__.env.remotePathTemplate
             .replaceAll('{model}', path_or_repo_id)
@@ -30188,14 +30460,14 @@ async function getModelFile(path_or_repo_id, filename, fatal = true, options = {
         filename
     );
 
-    // Choose cache key for filesystem cache
-    // When using the main revision (default), we use the request URL as the cache key.
-    // If a specific revision is requested, we account for this in the cache key.
-    let fsCacheKey = revision === 'main' ? requestURL : pathJoin(path_or_repo_id, revision, filename);
-
     /** @type {string} */
     let cacheKey;
-    let proposedCacheKey = cache instanceof FileCache ? fsCacheKey : remoteURL;
+    const proposedCacheKey = cache instanceof FileCache
+        // Choose cache key for filesystem cache
+        // When using the main revision (default), we use the request URL as the cache key.
+        // If a specific revision is requested, we account for this in the cache key.
+        ? revision === 'main' ? requestURL : pathJoin(path_or_repo_id, revision, filename)
+        : remoteURL;
 
     // Whether to cache the final response in the end.
     let toCacheResponse = false;
@@ -30208,11 +30480,10 @@ async function getModelFile(path_or_repo_id, filename, fatal = true, options = {
         //  1. We first try to get from cache using the local path. In some environments (like deno),
         //     non-URL cache keys are not allowed. In these cases, `response` will be undefined.
         //  2. If no response is found, we try to get from cache using the remote URL or file system cache.
-        response = await tryCache(cache, cachePath, proposedCacheKey);
+        response = await tryCache(cache, localPath, proposedCacheKey);
     }
 
     const cacheHit = response !== undefined;
-
     if (response === undefined) {
         // Caching not available, or file is not cached, so we perform the request
 
@@ -30230,9 +30501,9 @@ async function getModelFile(path_or_repo_id, filename, fatal = true, options = {
                     console.warn(`Unable to load from local path "${localPath}": "${e}"`);
                 }
             } else if (options.local_files_only) {
-                throw new Error(`\`local_files_only=true\`, but attempted to load a remote file from: ${localPath}.`);
+                throw new Error(`\`local_files_only=true\`, but attempted to load a remote file from: ${requestURL}.`);
             } else if (!_env_js__WEBPACK_IMPORTED_MODULE_2__.env.allowRemoteModels) {
-                throw new Error(`\`env.allowRemoteModels=false\`, but attempted to load a remote file from: ${localPath}.`);
+                throw new Error(`\`env.allowRemoteModels=false\`, but attempted to load a remote file from: ${requestURL}.`);
             }
         }
 
@@ -30252,6 +30523,11 @@ async function getModelFile(path_or_repo_id, filename, fatal = true, options = {
                     return null;
                 }
             }
+            if (!validModelId) {
+                // Before making any requests to the remote server, we check if the model ID is valid.
+                // This prevents unnecessary network requests for invalid model IDs.
+                throw Error(`Local file missing at "${localPath}" and download aborted due to invalid model ID "${path_or_repo_id}".`);
+            }
 
             // File not found locally, so we try to download it from the remote server
             response = await getFile(remoteURL);
@@ -30583,7 +30859,7 @@ class RawImage {
 
     /**
      * Helper method for reading an image from a variety of input types.
-     * @param {RawImage|string|URL} input
+     * @param {RawImage|string|URL|Blob|HTMLCanvasElement|OffscreenCanvas} input
      * @returns The image object.
      *
      * **Example:** Read image from a URL.
@@ -30602,6 +30878,14 @@ class RawImage {
             return input;
         } else if (typeof input === 'string' || input instanceof URL) {
             return await this.fromURL(input);
+        } else if (input instanceof Blob) {
+            return await this.fromBlob(input);
+        } else if (
+            (typeof HTMLCanvasElement !== "undefined" && input instanceof HTMLCanvasElement)
+            ||
+            (typeof OffscreenCanvas !== "undefined" && input instanceof OffscreenCanvas)
+        ) {
+            return this.fromCanvas(input);
         } else {
             throw new Error(`Unsupported input type: ${typeof input}`);
         }
@@ -33613,8 +33897,12 @@ function calc_unsqueeze_dims(dims, dim) {
  * @private
  */
 function safeIndex(index, size, dimension = null, boundsCheck = true) {
-    if (boundsCheck && (index < -size || index >= size)) {
-        throw new Error(`IndexError: index ${index} is out of bounds for dimension${dimension === null ? '' : ' ' + dimension} with size ${size}`);
+    if (index < -size || index >= size) {
+        if (boundsCheck) {
+            throw new Error(`IndexError: index ${index} is out of bounds for dimension${dimension === null ? '' : ' ' + dimension} with size ${size}`);
+        } else {
+            return index < -size ? 0 : size;
+        }
     }
 
     if (index < 0) {
@@ -34471,6 +34759,9 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony export */   Gemma2ForCausalLM: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Gemma2ForCausalLM),
 /* harmony export */   Gemma2Model: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Gemma2Model),
 /* harmony export */   Gemma2PreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Gemma2PreTrainedModel),
+/* harmony export */   Gemma3ForCausalLM: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Gemma3ForCausalLM),
+/* harmony export */   Gemma3Model: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Gemma3Model),
+/* harmony export */   Gemma3PreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Gemma3PreTrainedModel),
 /* harmony export */   GemmaForCausalLM: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.GemmaForCausalLM),
 /* harmony export */   GemmaModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.GemmaModel),
 /* harmony export */   GemmaPreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.GemmaPreTrainedModel),
@@ -34572,6 +34863,10 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony export */   MaskFormerPreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.MaskFormerPreTrainedModel),
 /* harmony export */   MaskedLMOutput: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.MaskedLMOutput),
 /* harmony export */   MaxLengthCriteria: () => (/* reexport safe */ _generation_stopping_criteria_js__WEBPACK_IMPORTED_MODULE_20__.MaxLengthCriteria),
+/* harmony export */   Metric3DForDepthEstimation: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Metric3DForDepthEstimation),
+/* harmony export */   Metric3DPreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Metric3DPreTrainedModel),
+/* harmony export */   Metric3Dv2ForDepthEstimation: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Metric3Dv2ForDepthEstimation),
+/* harmony export */   Metric3Dv2PreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Metric3Dv2PreTrainedModel),
 /* harmony export */   MgpstrForSceneTextRecognition: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.MgpstrForSceneTextRecognition),
 /* harmony export */   MgpstrModelOutput: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.MgpstrModelOutput),
 /* harmony export */   MgpstrPreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.MgpstrPreTrainedModel),
@@ -34724,11 +35019,19 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony export */   Qwen2VLImageProcessor: () => (/* reexport safe */ _models_image_processors_js__WEBPACK_IMPORTED_MODULE_14__.Qwen2VLImageProcessor),
 /* harmony export */   Qwen2VLPreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.Qwen2VLPreTrainedModel),
 /* harmony export */   Qwen2VLProcessor: () => (/* reexport safe */ _models_processors_js__WEBPACK_IMPORTED_MODULE_17__.Qwen2VLProcessor),
+/* harmony export */   RFDetrForObjectDetection: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RFDetrForObjectDetection),
+/* harmony export */   RFDetrModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RFDetrModel),
+/* harmony export */   RFDetrObjectDetectionOutput: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RFDetrObjectDetectionOutput),
+/* harmony export */   RFDetrPreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RFDetrPreTrainedModel),
 /* harmony export */   RTDetrForObjectDetection: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RTDetrForObjectDetection),
 /* harmony export */   RTDetrImageProcessor: () => (/* reexport safe */ _models_image_processors_js__WEBPACK_IMPORTED_MODULE_14__.RTDetrImageProcessor),
 /* harmony export */   RTDetrModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RTDetrModel),
 /* harmony export */   RTDetrObjectDetectionOutput: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RTDetrObjectDetectionOutput),
 /* harmony export */   RTDetrPreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RTDetrPreTrainedModel),
+/* harmony export */   RTDetrV2ForObjectDetection: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RTDetrV2ForObjectDetection),
+/* harmony export */   RTDetrV2Model: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RTDetrV2Model),
+/* harmony export */   RTDetrV2ObjectDetectionOutput: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RTDetrV2ObjectDetectionOutput),
+/* harmony export */   RTDetrV2PreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.RTDetrV2PreTrainedModel),
 /* harmony export */   RawAudio: () => (/* reexport safe */ _utils_audio_js__WEBPACK_IMPORTED_MODULE_5__.RawAudio),
 /* harmony export */   RawImage: () => (/* reexport safe */ _utils_image_js__WEBPACK_IMPORTED_MODULE_6__.RawImage),
 /* harmony export */   RawVideo: () => (/* reexport safe */ _utils_video_js__WEBPACK_IMPORTED_MODULE_7__.RawVideo),
@@ -34778,6 +35081,11 @@ __webpack_require__.r(__webpack_exports__);
 /* harmony export */   SmolVLMForConditionalGeneration: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.SmolVLMForConditionalGeneration),
 /* harmony export */   SmolVLMImageProcessor: () => (/* reexport safe */ _models_image_processors_js__WEBPACK_IMPORTED_MODULE_14__.SmolVLMImageProcessor),
 /* harmony export */   SmolVLMProcessor: () => (/* reexport safe */ _models_processors_js__WEBPACK_IMPORTED_MODULE_17__.SmolVLMProcessor),
+/* harmony export */   SnacDecoderModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.SnacDecoderModel),
+/* harmony export */   SnacEncoderModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.SnacEncoderModel),
+/* harmony export */   SnacFeatureExtractor: () => (/* reexport safe */ _models_feature_extractors_js__WEBPACK_IMPORTED_MODULE_11__.SnacFeatureExtractor),
+/* harmony export */   SnacModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.SnacModel),
+/* harmony export */   SnacPreTrainedModel: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.SnacPreTrainedModel),
 /* harmony export */   SpeechT5FeatureExtractor: () => (/* reexport safe */ _models_feature_extractors_js__WEBPACK_IMPORTED_MODULE_11__.SpeechT5FeatureExtractor),
 /* harmony export */   SpeechT5ForSpeechToText: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.SpeechT5ForSpeechToText),
 /* harmony export */   SpeechT5ForTextToSpeech: () => (/* reexport safe */ _models_js__WEBPACK_IMPORTED_MODULE_2__.SpeechT5ForTextToSpeech),