@huggingface/tasks 0.13.16 → 0.14.0

package/src/tasks/automatic-speech-recognition/spec/input.json

@@ -7,7 +7,8 @@
 	"properties": {
 		"inputs": {
 			"description": "The input audio data as a base64-encoded string. If no `parameters` are provided, you can also provide the audio data as a raw bytes payload.",
-			"type": "string"
+			"type": "string",
+			"comment": "type=binary"
 		},
 		"parameters": {
 			"description": "Additional inference parameters for Automatic Speech Recognition",

package/src/tasks/chat-completion/inference.ts

@@ -3,7 +3,6 @@
  *
  * Using src/scripts/inference-codegen
  */
-
 /**
  * Chat Completion Input.
  *
@@ -105,30 +104,24 @@ export interface ChatCompletionInput {
 	top_p?: number;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionInputMessage {
 	content: ChatCompletionInputMessageContent;
 	name?: string;
 	role: string;
 	[property: string]: unknown;
 }
-
 export type ChatCompletionInputMessageContent = ChatCompletionInputMessageChunk[] | string;
-
 export interface ChatCompletionInputMessageChunk {
 	image_url?: ChatCompletionInputURL;
 	text?: string;
 	type: ChatCompletionInputMessageChunkType;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionInputURL {
 	url: string;
 	[property: string]: unknown;
 }
-
 export type ChatCompletionInputMessageChunkType = "text" | "image_url";
-
 export interface ChatCompletionInputGrammarType {
 	type: ChatCompletionInputGrammarTypeType;
 	/**
@@ -140,9 +133,7 @@ export interface ChatCompletionInputGrammarType {
 	value: unknown;
 	[property: string]: unknown;
 }
-
 export type ChatCompletionInputGrammarTypeType = "json" | "regex";
-
 export interface ChatCompletionInputStreamOptions {
 	/**
 	 * If set, an additional chunk will be streamed before the data: [DONE] message. The usage
@@ -153,13 +144,11 @@ export interface ChatCompletionInputStreamOptions {
 	include_usage: boolean;
 	[property: string]: unknown;
 }
-
 /**
  *
  * <https://platform.openai.com/docs/guides/function-calling/configuring-function-calling-behavior-using-the-tool_choice-parameter>
  */
 export type ChatCompletionInputToolChoice = ChatCompletionInputToolChoiceEnum | ChatCompletionInputToolChoiceObject;
-
 /**
  * Means the model can pick between generating a message or calling one or more tools.
  *
@@ -168,30 +157,25 @@ export type ChatCompletionInputToolChoice = ChatCompletionInputToolChoiceEnum |
  * Means the model must call one or more tools.
  */
 export type ChatCompletionInputToolChoiceEnum = "auto" | "none" | "required";
-
 export interface ChatCompletionInputToolChoiceObject {
 	function: ChatCompletionInputFunctionName;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionInputFunctionName {
 	name: string;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionInputTool {
 	function: ChatCompletionInputFunctionDefinition;
 	type: string;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionInputFunctionDefinition {
 	arguments: unknown;
 	description?: string;
 	name: string;
 	[property: string]: unknown;
 }
-
 /**
  * Chat Completion Output.
  *
@@ -208,7 +192,6 @@ export interface ChatCompletionOutput {
 	usage: ChatCompletionOutputUsage;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionOutputComplete {
 	finish_reason: string;
 	index: number;
@@ -216,53 +199,45 @@ export interface ChatCompletionOutputComplete {
 	message: ChatCompletionOutputMessage;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionOutputLogprobs {
 	content: ChatCompletionOutputLogprob[];
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionOutputLogprob {
 	logprob: number;
 	token: string;
 	top_logprobs: ChatCompletionOutputTopLogprob[];
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionOutputTopLogprob {
 	logprob: number;
 	token: string;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionOutputMessage {
 	content?: string;
 	role: string;
 	tool_calls?: ChatCompletionOutputToolCall[];
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionOutputToolCall {
 	function: ChatCompletionOutputFunctionDefinition;
 	id: string;
 	type: string;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionOutputFunctionDefinition {
 	arguments: unknown;
 	description?: string;
 	name: string;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionOutputUsage {
 	completion_tokens: number;
 	prompt_tokens: number;
 	total_tokens: number;
 	[property: string]: unknown;
 }
-
 /**
  * Chat Completion Stream Output.
  *
@@ -279,7 +254,6 @@ export interface ChatCompletionStreamOutput {
 	usage?: ChatCompletionStreamOutputUsage;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionStreamOutputChoice {
 	delta: ChatCompletionStreamOutputDelta;
 	finish_reason?: string;
@@ -287,14 +261,12 @@ export interface ChatCompletionStreamOutputChoice {
 	logprobs?: ChatCompletionStreamOutputLogprobs;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionStreamOutputDelta {
 	content?: string;
 	role: string;
 	tool_calls?: ChatCompletionStreamOutputDeltaToolCall;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionStreamOutputDeltaToolCall {
 	function: ChatCompletionStreamOutputFunction;
 	id: string;
@@ -302,31 +274,26 @@ export interface ChatCompletionStreamOutputDeltaToolCall {
 	type: string;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionStreamOutputFunction {
 	arguments: string;
 	name?: string;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionStreamOutputLogprobs {
 	content: ChatCompletionStreamOutputLogprob[];
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionStreamOutputLogprob {
 	logprob: number;
 	token: string;
 	top_logprobs: ChatCompletionStreamOutputTopLogprob[];
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionStreamOutputTopLogprob {
 	logprob: number;
 	token: string;
 	[property: string]: unknown;
 }
-
 export interface ChatCompletionStreamOutputUsage {
 	completion_tokens: number;
 	prompt_tokens: number;

package/src/tasks/depth-estimation/inference.ts

@@ -3,7 +3,6 @@
  *
  * Using src/scripts/inference-codegen
  */
-
 /**
  * Inputs for Depth Estimation inference
  */
@@ -15,10 +14,11 @@ export interface DepthEstimationInput {
 	/**
 	 * Additional inference parameters for Depth Estimation
 	 */
-	parameters?: { [key: string]: unknown };
+	parameters?: {
+		[key: string]: unknown;
+	};
 	[property: string]: unknown;
 }
-
 /**
  * Outputs of inference for the Depth Estimation task
  */

package/src/tasks/document-question-answering/spec/input.json

@@ -11,7 +11,8 @@
 			"title": "DocumentQuestionAnsweringInputData",
 			"properties": {
 				"image": {
-					"description": "The image on which the question is asked"
+					"description": "The image on which the question is asked",
+					"comment": "type=binary"
 				},
 				"question": {
 					"type": "string",

package/src/tasks/feature-extraction/inference.ts

@@ -3,9 +3,7 @@
  *
  * Using src/scripts/inference-codegen
  */
-
 export type FeatureExtractionOutput = Array<number[]>;
-
 /**
  * Feature Extraction Input.
  *
@@ -36,5 +34,4 @@ export interface FeatureExtractionInput {
 	truncation_direction?: FeatureExtractionInputTruncationDirection;
 	[property: string]: unknown;
 }
-
 export type FeatureExtractionInputTruncationDirection = "Left" | "Right";

package/src/tasks/image-classification/inference.ts

@@ -11,7 +11,7 @@ export interface ImageClassificationInput {
 	 * The input image data as a base64-encoded string. If no `parameters` are provided, you can
 	 * also provide the image data as a raw bytes payload.
 	 */
-	inputs: string;
+	inputs: Blob;
 	/**
 	 * Additional inference parameters for Image Classification
 	 */

package/src/tasks/image-classification/spec/input.json

@@ -7,7 +7,8 @@
 	"properties": {
 		"inputs": {
 			"type": "string",
-			"description": "The input image data as a base64-encoded string. If no `parameters` are provided, you can also provide the image data as a raw bytes payload."
+			"description": "The input image data as a base64-encoded string. If no `parameters` are provided, you can also provide the image data as a raw bytes payload.",
+			"comment": "type=binary"
 		},
 		"parameters": {
 			"description": "Additional inference parameters for Image Classification",

package/src/tasks/image-segmentation/inference.ts

@@ -11,7 +11,7 @@ export interface ImageSegmentationInput {
 	 * The input image data as a base64-encoded string. If no `parameters` are provided, you can
 	 * also provide the image data as a raw bytes payload.
 	 */
-	inputs: string;
+	inputs: Blob;
 	/**
 	 * Additional inference parameters for Image Segmentation
 	 */

package/src/tasks/image-segmentation/spec/input.json

@@ -7,7 +7,8 @@
 	"properties": {
 		"inputs": {
 			"type": "string",
-			"description": "The input image data as a base64-encoded string. If no `parameters` are provided, you can also provide the image data as a raw bytes payload."
+			"description": "The input image data as a base64-encoded string. If no `parameters` are provided, you can also provide the image data as a raw bytes payload.",
+			"comment": "type=binary"
 		},
 		"parameters": {
 			"description": "Additional inference parameters for Image Segmentation",

package/src/tasks/image-to-image/inference.ts

@@ -3,7 +3,6 @@
  *
  * Using src/scripts/inference-codegen
  */
-
 /**
  * Inputs for Image To Image inference
  */
@@ -12,14 +11,13 @@ export interface ImageToImageInput {
 	 * The input image data as a base64-encoded string. If no `parameters` are provided, you can
 	 * also provide the image data as a raw bytes payload.
 	 */
-	inputs: string;
+	inputs: Blob;
 	/**
 	 * Additional inference parameters for Image To Image
 	 */
 	parameters?: ImageToImageParameters;
 	[property: string]: unknown;
 }
-
 /**
  * Additional inference parameters for Image To Image
  */
@@ -30,9 +28,9 @@ export interface ImageToImageParameters {
 	 */
 	guidance_scale?: number;
 	/**
-	 * One or several prompt to guide what NOT to include in image generation.
+	 * One prompt to guide what NOT to include in image generation.
 	 */
-	negative_prompt?: string[];
+	negative_prompt?: string;
 	/**
 	 * For diffusion models. The number of denoising steps. More denoising steps usually lead to
 	 * a higher quality image at the expense of slower inference.
@@ -44,7 +42,6 @@ export interface ImageToImageParameters {
 	target_size?: TargetSize;
 	[property: string]: unknown;
 }
-
 /**
  * The size in pixel of the output image.
  */
@@ -53,7 +50,6 @@ export interface TargetSize {
 	width: number;
 	[property: string]: unknown;
 }
-
 /**
  * Outputs of inference for the Image To Image task
  */

package/src/tasks/image-to-image/spec/input.json

@@ -7,7 +7,8 @@
 	"properties": {
 		"inputs": {
 			"type": "string",
-			"description": "The input image data as a base64-encoded string. If no `parameters` are provided, you can also provide the image data as a raw bytes payload."
+			"description": "The input image data as a base64-encoded string. If no `parameters` are provided, you can also provide the image data as a raw bytes payload.",
+			"comment": "type=binary"
 		},
 		"parameters": {
 			"description": "Additional inference parameters for Image To Image",
@@ -24,11 +25,8 @@
 					"description": "For diffusion models. A higher guidance scale value encourages the model to generate images closely linked to the text prompt at the expense of lower image quality."
 				},
 				"negative_prompt": {
-					"type": "array",
-					"items": {
-						"type": "string"
-					},
-					"description": "One or several prompt to guide what NOT to include in image generation."
+					"type": "string",
+					"description": "One prompt to guide what NOT to include in image generation."
 				},
 				"num_inference_steps": {
 					"type": "integer",

package/src/tasks/image-to-text/inference.ts

@@ -3,7 +3,6 @@
  *
  * Using src/scripts/inference-codegen
  */
-
 /**
  * Inputs for Image To Text inference
  */
@@ -11,14 +10,13 @@ export interface ImageToTextInput {
 	/**
 	 * The input image data
 	 */
-	inputs: unknown;
+	inputs: Blob;
 	/**
 	 * Additional inference parameters for Image To Text
 	 */
 	parameters?: ImageToTextParameters;
 	[property: string]: unknown;
 }
-
 /**
  * Additional inference parameters for Image To Text
  */
@@ -33,7 +31,6 @@ export interface ImageToTextParameters {
 	max_new_tokens?: number;
 	[property: string]: unknown;
 }
-
 /**
  * Parametrization of the text generation process
  */
@@ -120,12 +117,10 @@ export interface GenerationParameters {
 	use_cache?: boolean;
 	[property: string]: unknown;
 }
-
 /**
  * Controls the stopping condition for beam-based methods.
  */
 export type EarlyStoppingUnion = boolean | "never";
-
 /**
  * Outputs of inference for the Image To Text task
  */

package/src/tasks/image-to-text/spec/input.json

@@ -6,7 +6,8 @@
 	"type": "object",
 	"properties": {
 		"inputs": {
-			"description": "The input image data"
+			"description": "The input image data",
+			"comment": "type=binary"
 		},
 		"parameters": {
 			"description": "Additional inference parameters for Image To Text",

package/src/tasks/index.ts

@@ -73,6 +73,7 @@ export type * from "./sentence-similarity/inference.js";
 export type * from "./summarization/inference.js";
 export type * from "./table-question-answering/inference.js";
 export type { TextToImageInput, TextToImageOutput, TextToImageParameters } from "./text-to-image/inference.js";
+export type { TextToVideoParameters, TextToVideoOutput, TextToVideoInput } from "./text-to-video/inference.js";
 export type { TextToSpeechParameters, TextToSpeechInput, TextToSpeechOutput } from "./text-to-speech/inference.js";
 export type * from "./token-classification/inference.js";
 export type { TranslationInput, TranslationOutput } from "./translation/inference.js";

package/src/tasks/object-detection/inference.ts

@@ -11,7 +11,7 @@ export interface ObjectDetectionInput {
 	 * The input image data as a base64-encoded string. If no `parameters` are provided, you can
 	 * also provide the image data as a raw bytes payload.
 	 */
-	inputs: string;
+	inputs: Blob;
 	/**
 	 * Additional inference parameters for Object Detection
 	 */

package/src/tasks/object-detection/spec/input.json

@@ -7,7 +7,8 @@
 	"properties": {
 		"inputs": {
 			"type": "string",
-			"description": "The input image data as a base64-encoded string. If no `parameters` are provided, you can also provide the image data as a raw bytes payload."
+			"description": "The input image data as a base64-encoded string. If no `parameters` are provided, you can also provide the image data as a raw bytes payload.",
+			"comment": "type=binary"
 		},
 		"parameters": {
 			"description": "Additional inference parameters for Object Detection",

package/src/tasks/sentence-similarity/inference.ts

@@ -3,9 +3,7 @@
  *
  * Using src/scripts/inference-codegen
  */
-
 export type SentenceSimilarityOutput = number[];
-
 /**
  * Inputs for Sentence similarity inference
  */
@@ -14,10 +12,11 @@ export interface SentenceSimilarityInput {
 	/**
 	 * Additional inference parameters for Sentence Similarity
 	 */
-	parameters?: { [key: string]: unknown };
+	parameters?: {
+		[key: string]: unknown;
+	};
 	[property: string]: unknown;
 }
-
 export interface SentenceSimilarityInputData {
 	/**
 	 * A list of strings which will be compared against the source_sentence.

package/src/tasks/summarization/inference.ts

@@ -3,7 +3,6 @@
  *
  * Using src/scripts/inference-codegen
  */
-
 /**
  * Inputs for Summarization inference
  */
@@ -18,7 +17,6 @@ export interface SummarizationInput {
 	parameters?: SummarizationParameters;
 	[property: string]: unknown;
 }
-
 /**
  * Additional inference parameters for summarization.
  */
@@ -30,19 +28,19 @@ export interface SummarizationParameters {
 	/**
 	 * Additional parametrization of the text generation algorithm.
 	 */
-	generate_parameters?: { [key: string]: unknown };
+	generate_parameters?: {
+		[key: string]: unknown;
+	};
 	/**
 	 * The truncation strategy to use.
 	 */
 	truncation?: SummarizationTruncationStrategy;
 	[property: string]: unknown;
 }
-
 /**
  * The truncation strategy to use.
  */
 export type SummarizationTruncationStrategy = "do_not_truncate" | "longest_first" | "only_first" | "only_second";
-
 /**
  * Outputs of inference for the Summarization task
  */

package/src/tasks/text-generation/inference.ts

@@ -3,7 +3,6 @@
  *
  * Using src/scripts/inference-codegen
  */
-
 /**
  * Text Generation Input.
  *
@@ -17,7 +16,6 @@ export interface TextGenerationInput {
 	stream?: boolean;
 	[property: string]: unknown;
 }
-
 export interface TextGenerationInputGenerateParameters {
 	/**
 	 * Lora adapter id
@@ -100,7 +98,6 @@ export interface TextGenerationInputGenerateParameters {
 	watermark?: boolean;
 	[property: string]: unknown;
 }
-
 export interface TextGenerationInputGrammarType {
 	type: Type;
 	/**
@@ -112,9 +109,7 @@ export interface TextGenerationInputGrammarType {
 	value: unknown;
 	[property: string]: unknown;
 }
-
 export type Type = "json" | "regex";
-
 /**
  * Text Generation Output.
  *
@@ -127,7 +122,6 @@ export interface TextGenerationOutput {
 	generated_text: string;
 	[property: string]: unknown;
 }
-
 export interface TextGenerationOutputDetails {
 	best_of_sequences?: TextGenerationOutputBestOfSequence[];
 	finish_reason: TextGenerationOutputFinishReason;
@@ -138,7 +132,6 @@ export interface TextGenerationOutputDetails {
 	top_tokens?: Array<TextGenerationOutputToken[]>;
 	[property: string]: unknown;
 }
-
 export interface TextGenerationOutputBestOfSequence {
 	finish_reason: TextGenerationOutputFinishReason;
 	generated_text: string;
@@ -149,16 +142,13 @@ export interface TextGenerationOutputBestOfSequence {
 	top_tokens?: Array<TextGenerationOutputToken[]>;
 	[property: string]: unknown;
 }
-
 export type TextGenerationOutputFinishReason = "length" | "eos_token" | "stop_sequence";
-
 export interface TextGenerationOutputPrefillToken {
 	id: number;
 	logprob: number;
 	text: string;
 	[property: string]: unknown;
 }
-
 export interface TextGenerationOutputToken {
 	id: number;
 	logprob: number;
@@ -166,7 +156,6 @@ export interface TextGenerationOutputToken {
 	text: string;
 	[property: string]: unknown;
 }
-
 /**
  * Text Generation Stream Output.
  *
@@ -182,7 +171,6 @@ export interface TextGenerationStreamOutput {
 	top_tokens?: TextGenerationStreamOutputToken[];
 	[property: string]: unknown;
 }
-
 export interface TextGenerationStreamOutputStreamDetails {
 	finish_reason: TextGenerationOutputFinishReason;
 	generated_tokens: number;
@@ -190,7 +178,6 @@ export interface TextGenerationStreamOutputStreamDetails {
 	seed?: number;
 	[property: string]: unknown;
 }
-
 export interface TextGenerationStreamOutputToken {
 	id: number;
 	logprob: number;

package/src/tasks/text-to-audio/inference.ts

@@ -1,9 +1,22 @@
+/**
+ * Outputs of inference for the Text To Audio task
+ */
+export interface TextToAudioOutput {
+	/**
+	 * The generated audio waveform.
+	 */
+	audio: Blob;
+	/**
+	 * The sampling rate of the generated audio waveform.
+	 */
+	sampling_rate: number;
+	[property: string]: unknown;
+}
 /**
  * Inference code generated from the JSON schema spec in ./spec
  *
  * Using src/scripts/inference-codegen
  */
-
 /**
  * Inputs for Text To Audio inference
  */
@@ -18,7 +31,6 @@ export interface TextToAudioInput {
 	parameters?: TextToAudioParameters;
 	[property: string]: unknown;
 }
-
 /**
  * Additional inference parameters for Text To Audio
  */
@@ -29,7 +41,6 @@ export interface TextToAudioParameters {
 	generation_parameters?: GenerationParameters;
 	[property: string]: unknown;
 }
-
 /**
  * Parametrization of the text generation process
  */
@@ -116,24 +127,7 @@ export interface GenerationParameters {
 	use_cache?: boolean;
 	[property: string]: unknown;
 }
-
 /**
  * Controls the stopping condition for beam-based methods.
  */
 export type EarlyStoppingUnion = boolean | "never";
-
-/**
- * Outputs of inference for the Text To Audio task
- */
-export interface TextToAudioOutput {
-	/**
-	 * The generated audio waveform.
-	 */
-	audio: unknown;
-	samplingRate: unknown;
-	/**
-	 * The sampling rate of the generated audio waveform.
-	 */
-	sampling_rate?: number;
-	[property: string]: unknown;
-}