@huggingface/tasks 0.3.3 → 0.4.0

package/dist/index.js

@@ -1260,7 +1260,7 @@ var PIPELINE_DATA = {
     color: "green"
   },
   "image-text-to-text": {
-    name: "Image + Text to Text (VLLMs)",
+    name: "Image-Text-to-Text",
     modality: "multimodal",
     color: "red",
     hideInDatasets: true
@@ -4187,7 +4187,7 @@ model = joblib.load(
 };
 var sklearn = (model) => {
   if (model.tags?.includes("skops")) {
-    const skopsmodelFile = model.config?.sklearn?.filename;
+    const skopsmodelFile = model.config?.sklearn?.model?.file;
     const skopssaveFormat = model.config?.sklearn?.model_format;
     if (!skopsmodelFile) {
       return [`# \u26A0\uFE0F Model filename not specified in config.json`];
@@ -4257,7 +4257,7 @@ var speechBrainMethod = (speechbrainInterface) => {
   }
 };
 var speechbrain = (model) => {
-  const speechbrainInterface = model.config?.speechbrain?.interface;
+  const speechbrainInterface = model.config?.speechbrain?.speechbrain_interface;
   if (speechbrainInterface === void 0) {
     return [`# interface not specified in config.json`];
   }
@@ -4335,7 +4335,7 @@ var peftTask = (peftTaskType) => {
   }
 };
 var peft = (model) => {
-  const { base_model_name: peftBaseModel, task_type: peftTaskType } = model.config?.peft ?? {};
+  const { base_model_name_or_path: peftBaseModel, task_type: peftTaskType } = model.config?.peft ?? {};
   const pefttask = peftTask(peftTaskType);
   if (!pefttask) {
     return [`Task type is invalid.`];
@@ -4406,6 +4406,41 @@ var pythae = (model) => [
 
 model = AutoModel.load_from_hf_hub("${model.id}")`
 ];
+var musicgen = (model) => [
+  `from audiocraft.models import MusicGen
+
+model = MusicGen.get_pretrained("${model.id}")
+
+descriptions = ['happy rock', 'energetic EDM', 'sad jazz']
+wav = model.generate(descriptions)  # generates 3 samples.`
+];
+var magnet = (model) => [
+  `from audiocraft.models import MAGNeT
+	
+model = MAGNeT.get_pretrained("${model.id}")
+
+descriptions = ['disco beat', 'energetic EDM', 'funky groove']
+wav = model.generate(descriptions)  # generates 3 samples.`
+];
+var audiogen = (model) => [
+  `from audiocraft.models import AudioGen
+	
+model = AudioGen.get_pretrained("${model.id}")
+model.set_generation_params(duration=5)  # generate 5 seconds.
+descriptions = ['dog barking', 'sirene of an emergency vehicle', 'footsteps in a corridor']
+wav = model.generate(descriptions)  # generates 3 samples.`
+];
+var audiocraft = (model) => {
+  if (model.tags?.includes("musicgen")) {
+    return musicgen(model);
+  } else if (model.tags?.includes("audiogen")) {
+    return audiogen(model);
+  } else if (model.tags?.includes("magnet")) {
+    return magnet(model);
+  } else {
+    return [`# Type of model unknown.`];
+  }
+};
 
 // src/model-libraries.ts
 var MODEL_LIBRARIES_UI_ELEMENTS = {
@@ -4439,6 +4474,13 @@ var MODEL_LIBRARIES_UI_ELEMENTS = {
       term: { path: "pytorch_model.bin" }
     }
   },
+  audiocraft: {
+    prettyLabel: "Audiocraft",
+    repoName: "audiocraft",
+    repoUrl: "https://github.com/facebookresearch/audiocraft",
+    snippets: audiocraft,
+    filter: false
+  },
   bertopic: {
     prettyLabel: "BERTopic",
     repoName: "BERTopic",

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@huggingface/tasks",
   "packageManager": "pnpm@8.10.5",
-  "version": "0.3.3",
+  "version": "0.4.0",
   "description": "List of ML tasks for huggingface.co/tasks",
   "repository": "https://github.com/huggingface/huggingface.js.git",
   "publishConfig": {

package/src/index.ts

@@ -19,12 +19,14 @@ export type { LibraryUiElement, ModelLibraryKey } from "./model-libraries";
 export type { ModelData, TransformersInfo } from "./model-data";
 export type { SpecialTokensMap, TokenizerConfig } from "./tokenizer-data";
 export type {
+	ChatMessage,
 	WidgetExample,
 	WidgetExampleAttribute,
 	WidgetExampleAssetAndPromptInput,
 	WidgetExampleAssetAndTextInput,
 	WidgetExampleAssetAndZeroShotInput,
 	WidgetExampleAssetInput,
+	WidgetExampleChatInput,
 	WidgetExampleSentenceSimilarityInput,
 	WidgetExampleStructuredDataInput,
 	WidgetExampleTableDataInput,

package/src/model-data.ts

@@ -40,21 +40,46 @@ export interface ModelData {
 	/**
 	 * this dictionary has useful information about the model configuration
 	 */
-	config?: Record<string, unknown> & {
-		adapter_transformers?: { model_class?: string; model_name?: string };
+	config?: {
 		architectures?: string[];
+		/**
+		 * Dict of AutoModel or Auto… class name to local import path in the repo
+		 */
+		auto_map?: {
+			/**
+			 * String Property
+			 */
+			[x: string]: string;
+		};
+		model_type?: string;
+		quantization_config?: {
+			bits?: number;
+			load_in_4bit?: boolean;
+			load_in_8bit?: boolean;
+		};
+		tokenizer_config?: TokenizerConfig;
+		adapter_transformers?: {
+			model_name?: string;
+			model_class?: string;
+		};
+		diffusers?: {
+			_class_name?: string;
+		};
 		sklearn?: {
-			filename?: string;
+			model?: {
+				file?: string;
+			};
 			model_format?: string;
 		};
 		speechbrain?: {
-			interface?: string;
+			speechbrain_interface?: string;
+			vocoder_interface?: string;
+			vocoder_model_id?: string;
 		};
 		peft?: {
-			base_model_name?: string;
+			base_model_name_or_path?: string;
 			task_type?: string;
 		};
-		tokenizer_config?: TokenizerConfig;
 	};
 	/**
 	 * all the model tags
@@ -80,7 +105,7 @@ export interface ModelData {
 	 */
 	widgetData?: WidgetExample[] | undefined;
 	/**
-	 * Parameters that will be used by the widget when calling Inference Endpoints (serverless)
+	 * Parameters that will be used by the widget when calling Inference API (serverless)
 	 * https://huggingface.co/docs/api-inference/detailed_parameters
 	 *
 	 * can be set in the model card metadata (under `inference/parameters`)

package/src/model-libraries-snippets.ts

@@ -293,7 +293,7 @@ model = joblib.load(
 
 export const sklearn = (model: ModelData): string[] => {
 	if (model.tags?.includes("skops")) {
-		const skopsmodelFile = model.config?.sklearn?.filename;
+		const skopsmodelFile = model.config?.sklearn?.model?.file;
 		const skopssaveFormat = model.config?.sklearn?.model_format;
 		if (!skopsmodelFile) {
 			return [`# ⚠️ Model filename not specified in config.json`];
@@ -372,7 +372,7 @@ const speechBrainMethod = (speechbrainInterface: string) => {
 };
 
 export const speechbrain = (model: ModelData): string[] => {
-	const speechbrainInterface = model.config?.speechbrain?.interface;
+	const speechbrainInterface = model.config?.speechbrain?.speechbrain_interface;
 	if (speechbrainInterface === undefined) {
 		return [`# interface not specified in config.json`];
 	}
@@ -465,7 +465,7 @@ const peftTask = (peftTaskType?: string) => {
 };
 
 export const peft = (model: ModelData): string[] => {
-	const { base_model_name: peftBaseModel, task_type: peftTaskType } = model.config?.peft ?? {};
+	const { base_model_name_or_path: peftBaseModel, task_type: peftTaskType } = model.config?.peft ?? {};
 	const pefttask = peftTask(peftTaskType);
 	if (!pefttask) {
 		return [`Task type is invalid.`];
@@ -548,4 +548,42 @@ export const pythae = (model: ModelData): string[] => [
 model = AutoModel.load_from_hf_hub("${model.id}")`,
 ];
 
+const musicgen = (model: ModelData): string[] => [
+	`from audiocraft.models import MusicGen
+
+model = MusicGen.get_pretrained("${model.id}")
+
+descriptions = ['happy rock', 'energetic EDM', 'sad jazz']
+wav = model.generate(descriptions)  # generates 3 samples.`,
+];
+
+const magnet = (model: ModelData): string[] => [
+	`from audiocraft.models import MAGNeT
+	
+model = MAGNeT.get_pretrained("${model.id}")
+
+descriptions = ['disco beat', 'energetic EDM', 'funky groove']
+wav = model.generate(descriptions)  # generates 3 samples.`,
+];
+
+const audiogen = (model: ModelData): string[] => [
+	`from audiocraft.models import AudioGen
+	
+model = AudioGen.get_pretrained("${model.id}")
+model.set_generation_params(duration=5)  # generate 5 seconds.
+descriptions = ['dog barking', 'sirene of an emergency vehicle', 'footsteps in a corridor']
+wav = model.generate(descriptions)  # generates 3 samples.`,
+];
+
+export const audiocraft = (model: ModelData): string[] => {
+	if (model.tags?.includes("musicgen")) {
+		return musicgen(model);
+	} else if (model.tags?.includes("audiogen")) {
+		return audiogen(model);
+	} else if (model.tags?.includes("magnet")) {
+		return magnet(model);
+	} else {
+		return [`# Type of model unknown.`];
+	}
+};
 //#endregion

package/src/model-libraries.ts

@@ -88,6 +88,13 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
 			term: { path: "pytorch_model.bin" },
 		},
 	},
+	audiocraft: {
+		prettyLabel: "Audiocraft",
+		repoName: "audiocraft",
+		repoUrl: "https://github.com/facebookresearch/audiocraft",
+		snippets: snippets.audiocraft,
+		filter: false,
+	},
 	bertopic: {
 		prettyLabel: "BERTopic",
 		repoName: "BERTopic",

package/src/pipelines.ts

@@ -590,7 +590,7 @@ export const PIPELINE_DATA = {
 		color: "green",
 	},
 	"image-text-to-text": {
-		name: "Image + Text to Text (VLLMs)",
+		name: "Image-Text-to-Text",
 		modality: "multimodal",
 		color: "red",
 		hideInDatasets: true,

package/src/tasks/index.ts

@@ -36,6 +36,64 @@ import zeroShotClassification from "./zero-shot-classification/data";
 import zeroShotImageClassification from "./zero-shot-image-classification/data";
 import zeroShotObjectDetection from "./zero-shot-object-detection/data";
 
+export type * from "./audio-classification/inference";
+export type * from "./automatic-speech-recognition/inference";
+export type * from "./document-question-answering/inference";
+export type * from "./feature-extraction/inference";
+export type * from "./fill-mask/inference";
+export type {
+	ImageClassificationInput,
+	ImageClassificationOutput,
+	ImageClassificationOutputElement,
+	ImageClassificationParameters,
+} from "./image-classification/inference";
+export type * from "./image-to-image/inference";
+export type { ImageToTextInput, ImageToTextOutput, ImageToTextParameters } from "./image-to-text/inference";
+export type * from "./image-segmentation/inference";
+export type * from "./object-detection/inference";
+export type * from "./depth-estimation/inference";
+export type * from "./question-answering/inference";
+export type * from "./sentence-similarity/inference";
+export type * from "./summarization/inference";
+export type * from "./table-question-answering/inference";
+export type { TextToImageInput, TextToImageOutput, TextToImageParameters } from "./text-to-image/inference";
+export type { TextToAudioParameters, TextToSpeechInput, TextToSpeechOutput } from "./text-to-speech/inference";
+export type * from "./token-classification/inference";
+export type {
+	Text2TextGenerationParameters,
+	Text2TextGenerationTruncationStrategy,
+	TranslationInput,
+	TranslationOutput,
+} from "./translation/inference";
+export type {
+	ClassificationOutputTransform,
+	TextClassificationInput,
+	TextClassificationOutput,
+	TextClassificationOutputElement,
+	TextClassificationParameters,
+} from "./text-classification/inference";
+export type {
+	FinishReason,
+	PrefillToken,
+	TextGenerationInput,
+	TextGenerationOutput,
+	TextGenerationOutputDetails,
+	TextGenerationParameters,
+	TextGenerationSequenceDetails,
+	Token,
+} from "./text-generation/inference";
+export type * from "./video-classification/inference";
+export type * from "./visual-question-answering/inference";
+export type * from "./zero-shot-classification/inference";
+export type * from "./zero-shot-image-classification/inference";
+export type {
+	BoundingBox,
+	ZeroShotObjectDetectionInput,
+	ZeroShotObjectDetectionInputData,
+	ZeroShotObjectDetectionOutput,
+	ZeroShotObjectDetectionOutputElement,
+} from "./zero-shot-object-detection/inference";
+
 import type { ModelLibraryKey } from "../model-libraries";
 
 /**

package/src/tasks/summarization/inference.ts

@@ -45,15 +45,12 @@ export interface Text2TextGenerationParameters {
 export type Text2TextGenerationTruncationStrategy = "do_not_truncate" | "longest_first" | "only_first" | "only_second";
 
 /**
- * Outputs for Summarization inference
- *
- * Outputs of inference for the Text2text Generation task
+ * Outputs of inference for the Summarization task
  */
 export interface SummarizationOutput {
-	generatedText: unknown;
 	/**
-	 * The generated text.
+	 * The summarized text.
 	 */
-	generated_text?: string;
+	summary_text: string;
 	[property: string]: unknown;
 }

package/src/tasks/summarization/spec/output.json

@@ -1,7 +1,14 @@
 {
-	"$ref": "/inference/schemas/text2text-generation/output.json",
 	"$id": "/inference/schemas/summarization/output.json",
 	"$schema": "http://json-schema.org/draft-06/schema#",
+	"description": "Outputs of inference for the Summarization task",
 	"title": "SummarizationOutput",
-	"description": "Outputs for Summarization inference"
+	"type": "object",
+	"properties": {
+		"summary_text": {
+			"type": "string",
+			"description": "The summarized text."
+		}
+	},
+	"required": ["summary_text"]
 }

package/src/tasks/translation/inference.ts

@@ -45,15 +45,12 @@ export interface Text2TextGenerationParameters {
 export type Text2TextGenerationTruncationStrategy = "do_not_truncate" | "longest_first" | "only_first" | "only_second";
 
 /**
- * Outputs for Translation inference
- *
- * Outputs of inference for the Text2text Generation task
+ * Outputs of inference for the Translation task
  */
 export interface TranslationOutput {
-	generatedText: unknown;
 	/**
-	 * The generated text.
+	 * The translated text.
 	 */
-	generated_text?: string;
+	translation_text: string;
 	[property: string]: unknown;
 }

package/src/tasks/translation/spec/output.json

@@ -1,7 +1,14 @@
 {
-	"$ref": "/inference/schemas/text2text-generation/output.json",
 	"$id": "/inference/schemas/translation/output.json",
 	"$schema": "http://json-schema.org/draft-06/schema#",
+	"description": "Outputs of inference for the Translation task",
 	"title": "TranslationOutput",
-	"description": "Outputs for Translation inference"
+	"type": "object",
+	"properties": {
+		"translation_text": {
+			"type": "string",
+			"description": "The translated text."
+		}
+	},
+	"required": ["translation_text"]
 }

package/src/widget-example.ts

@@ -51,6 +51,15 @@ export interface WidgetExampleBase<TOutput> {
 	output?: TOutput;
 }
 
+export interface ChatMessage {
+	role: "user" | "assistant" | "system";
+	content: string;
+}
+
+export interface WidgetExampleChatInput<TOutput = WidgetExampleOutput> extends WidgetExampleBase<TOutput> {
+	messages: ChatMessage[];
+}
+
 export interface WidgetExampleTextInput<TOutput = WidgetExampleOutput> extends WidgetExampleBase<TOutput> {
 	text: string;
 }
@@ -101,6 +110,7 @@ export interface WidgetExampleSentenceSimilarityInput<TOutput = WidgetExampleOut
 //#endregion
 
 export type WidgetExample<TOutput = WidgetExampleOutput> =
+	| WidgetExampleChatInput<TOutput>
 	| WidgetExampleTextInput<TOutput>
 	| WidgetExampleTextAndContextInput<TOutput>
 	| WidgetExampleTextAndTableInput<TOutput>