@huggingface/transformers 3.4.0 → 3.4.1

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@huggingface/transformers",
-  "version": "3.4.0",
+  "version": "3.4.1",
   "description": "State-of-the-art Machine Learning for the web. Run 🤗 Transformers directly in your browser, with no need for a server!",
   "main": "./src/transformers.js",
   "types": "./types/transformers.d.ts",

package/src/configs.js

@@ -67,6 +67,7 @@ function getNormalizedConfig(config) {
         // Sub-configs
         case 'llava':
         case 'paligemma':
+        case 'gemma3':
         case 'florence2':
         case 'llava_onevision':
         case 'idefics3':
@@ -126,6 +127,7 @@ function getNormalizedConfig(config) {
             break;
         case 'gemma':
         case 'gemma2':
+        case 'gemma3_text':
         case 'glm':
         case 'helium':
             mapping['num_heads'] = 'num_key_value_heads';

package/src/env.js

@@ -26,7 +26,7 @@ import fs from 'fs';
 import path from 'path';
 import url from 'url';
 
-const VERSION = '3.4.0';
+const VERSION = '3.4.1';
 
 // Check if various APIs are available (depends on environment)
 const IS_BROWSER_ENV = typeof window !== "undefined" && typeof window.document !== "undefined";

package/src/models/feature_extractors.js

@@ -6,6 +6,7 @@ export * from './dac/feature_extraction_dac.js';
 export * from './moonshine/feature_extraction_moonshine.js';
 export * from './pyannote/feature_extraction_pyannote.js';
 export * from './seamless_m4t/feature_extraction_seamless_m4t.js';
+export * from './snac/feature_extraction_snac.js';
 export * from './speecht5/feature_extraction_speecht5.js';
 export * from './wav2vec2/feature_extraction_wav2vec2.js';
 export * from './wespeaker/feature_extraction_wespeaker.js';

package/src/models/snac/feature_extraction_snac.js

@@ -0,0 +1,3 @@
+import { DacFeatureExtractor } from '../dac/feature_extraction_dac.js';
+
+export class SnacFeatureExtractor extends DacFeatureExtractor { }

package/src/models.js

@@ -594,8 +594,8 @@ 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);
     }
 
@@ -4520,6 +4520,23 @@ export class Gemma2Model extends Gemma2PreTrainedModel { }
 export class Gemma2ForCausalLM extends Gemma2PreTrainedModel { }
 //////////////////////////////////////////////////
 
+
+//////////////////////////////////////////////////
+// Gemma3 models
+
+/**
+ * The bare Gemma3 Model outputting raw hidden-states without any specific head on top.
+ */
+export class Gemma3PreTrainedModel extends PreTrainedModel { }
+/**
+ * The bare Gemma3 Model outputting raw hidden-states without any specific head on top.
+ */
+export class Gemma3Model extends Gemma3PreTrainedModel { }
+
+export class Gemma3ForCausalLM extends Gemma3PreTrainedModel { }
+//////////////////////////////////////////////////
+
+
 //////////////////////////////////////////////////
 export class OpenELMPreTrainedModel extends PreTrainedModel { }
 export class OpenELMModel extends OpenELMPreTrainedModel { }
@@ -5372,6 +5389,16 @@ export class DepthProPreTrainedModel extends PreTrainedModel { }
 export class DepthProForDepthEstimation extends DepthProPreTrainedModel { }
 //////////////////////////////////////////////////
 
+//////////////////////////////////////////////////
+export class Metric3DPreTrainedModel extends PreTrainedModel { }
+export class Metric3DForDepthEstimation extends Metric3DPreTrainedModel { }
+//////////////////////////////////////////////////
+
+//////////////////////////////////////////////////
+export class Metric3Dv2PreTrainedModel extends PreTrainedModel { }
+export class Metric3Dv2ForDepthEstimation extends Metric3Dv2PreTrainedModel { }
+//////////////////////////////////////////////////
+
 //////////////////////////////////////////////////
 export class MaskFormerPreTrainedModel extends PreTrainedModel { }
 export class MaskFormerModel extends MaskFormerPreTrainedModel { }
@@ -7287,6 +7314,60 @@ export class DacDecoderModel extends DacPreTrainedModel {
 }
 //////////////////////////////////////////////////
 
+
+//////////////////////////////////////////////////
+// Snac models
+export class SnacPreTrainedModel extends PreTrainedModel {
+    main_input_name = 'input_values';
+    forward_params = ['input_values'];
+}
+
+/**
+ * The SNAC (Multi-Scale Neural Audio Codec) model.
+ */
+export 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);
+    }
+}
+
+export 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',
+        });
+    }
+}
+export 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)
@@ -7468,6 +7549,7 @@ 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([
@@ -7488,6 +7570,7 @@ 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]],
@@ -7587,6 +7670,7 @@ 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]],
@@ -7788,6 +7872,8 @@ 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([
@@ -7873,6 +7959,8 @@ 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);