@huggingface/tasks 0.12.7 → 0.12.9

package/src/tasks/chat-completion/spec/stream_output.json

@@ -4,7 +4,7 @@
 	"description": "Chat Completion Stream Output.\n\nAuto-generated from TGI specs.\nFor more details, check out https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/scripts/inference-tgi-import.ts.",
 	"title": "ChatCompletionStreamOutput",
 	"type": "object",
-	"required": ["id", "object", "created", "model", "system_fingerprint", "choices"],
+	"required": ["id", "created", "model", "system_fingerprint", "choices"],
 	"properties": {
 		"choices": {
 			"type": "array",
@@ -25,11 +25,16 @@
 			"type": "string",
 			"example": "mistralai/Mistral-7B-Instruct-v0.2"
 		},
-		"object": {
-			"type": "string"
-		},
 		"system_fingerprint": {
 			"type": "string"
+		},
+		"usage": {
+			"allOf": [
+				{
+					"$ref": "#/$defs/ChatCompletionStreamOutputUsage"
+				}
+			],
+			"nullable": true
 		}
 	},
 	"$defs": {
@@ -61,28 +66,44 @@
 			"title": "ChatCompletionStreamOutputChoice"
 		},
 		"ChatCompletionStreamOutputDelta": {
+			"oneOf": [
+				{
+					"$ref": "#/$defs/ChatCompletionStreamOutputTextMessage"
+				},
+				{
+					"$ref": "#/$defs/ChatCompletionStreamOutputToolCallDelta"
+				}
+			],
+			"title": "ChatCompletionStreamOutputDelta"
+		},
+		"ChatCompletionStreamOutputTextMessage": {
 			"type": "object",
-			"required": ["role"],
+			"required": ["role", "content"],
 			"properties": {
 				"content": {
 					"type": "string",
-					"example": "What is Deep Learning?",
-					"nullable": true
+					"example": "My name is David and I"
 				},
 				"role": {
 					"type": "string",
 					"example": "user"
+				}
+			},
+			"title": "ChatCompletionStreamOutputTextMessage"
+		},
+		"ChatCompletionStreamOutputToolCallDelta": {
+			"type": "object",
+			"required": ["role", "tool_calls"],
+			"properties": {
+				"role": {
+					"type": "string",
+					"example": "assistant"
 				},
 				"tool_calls": {
-					"allOf": [
-						{
-							"$ref": "#/$defs/ChatCompletionStreamOutputDeltaToolCall"
-						}
-					],
-					"nullable": true
+					"$ref": "#/$defs/ChatCompletionStreamOutputDeltaToolCall"
 				}
 			},
-			"title": "ChatCompletionStreamOutputDelta"
+			"title": "ChatCompletionStreamOutputToolCallDelta"
 		},
 		"ChatCompletionStreamOutputDeltaToolCall": {
 			"type": "object",
@@ -165,6 +186,28 @@
 				}
 			},
 			"title": "ChatCompletionStreamOutputTopLogprob"
+		},
+		"ChatCompletionStreamOutputUsage": {
+			"type": "object",
+			"required": ["prompt_tokens", "completion_tokens", "total_tokens"],
+			"properties": {
+				"completion_tokens": {
+					"type": "integer",
+					"format": "int32",
+					"minimum": 0
+				},
+				"prompt_tokens": {
+					"type": "integer",
+					"format": "int32",
+					"minimum": 0
+				},
+				"total_tokens": {
+					"type": "integer",
+					"format": "int32",
+					"minimum": 0
+				}
+			},
+			"title": "ChatCompletionStreamOutputUsage"
 		}
 	}
 }

package/src/tasks/common-definitions.json

@@ -59,7 +59,7 @@
 				},
 				"max_new_tokens": {
 					"type": "integer",
-					"description": "The maximum number of tokens to generate. Takes precedence over maxLength."
+					"description": "The maximum number of tokens to generate. Takes precedence over max_length."
 				},
 				"min_length": {
 					"type": "integer",
@@ -67,7 +67,7 @@
 				},
 				"min_new_tokens": {
 					"type": "integer",
-					"description": "The minimum number of tokens to generate. Takes precedence over maxLength."
+					"description": "The minimum number of tokens to generate. Takes precedence over min_length."
 				},
 				"do_sample": {
 					"type": "boolean",

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

@@ -72,7 +72,7 @@ export interface GenerationParameters {
 	 */
 	max_length?: number;
 	/**
-	 * The maximum number of tokens to generate. Takes precedence over maxLength.
+	 * The maximum number of tokens to generate. Takes precedence over max_length.
 	 */
 	max_new_tokens?: number;
 	/**
@@ -80,7 +80,7 @@ export interface GenerationParameters {
 	 */
 	min_length?: number;
 	/**
-	 * The minimum number of tokens to generate. Takes precedence over maxLength.
+	 * The minimum number of tokens to generate. Takes precedence over min_length.
 	 */
 	min_new_tokens?: number;
 	/**

package/src/tasks/index.ts

@@ -39,6 +39,7 @@ import zeroShotImageClassification from "./zero-shot-image-classification/data";
 import zeroShotObjectDetection from "./zero-shot-object-detection/data";
 import imageTo3D from "./image-to-3d/data";
 import textTo3D from "./text-to-3d/data";
+import keypointDetection from "./keypoint-detection/data";
 
 export type * from "./audio-classification/inference";
 export type * from "./automatic-speech-recognition/inference";
@@ -71,7 +72,7 @@ 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 { TextToSpeechParameters, TextToSpeechInput, TextToSpeechOutput } from "./text-to-speech/inference";
 export type * from "./token-classification/inference";
 export type { TranslationInput, TranslationOutput } from "./translation/inference";
 export type {
@@ -208,7 +209,7 @@ export const TASKS_DATA: Record<PipelineType, TaskData | undefined> = {
 	"image-text-to-text": getData("image-text-to-text", imageTextToText),
 	"image-to-text": getData("image-to-text", imageToText),
 	"image-to-video": undefined,
-	"keypoint-detection": getData("keypoint-detection", placeholder),
+	"keypoint-detection": getData("keypoint-detection", keypointDetection),
 	"mask-generation": getData("mask-generation", maskGeneration),
 	"multiple-choice": undefined,
 	"object-detection": getData("object-detection", objectDetection),

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

@@ -19,23 +19,84 @@ export interface TextGenerationInput {
 }
 
 export interface TextGenerationInputGenerateParameters {
+	/**
+	 * Lora adapter id
+	 */
+	adapter_id?: string;
+	/**
+	 * Generate best_of sequences and return the one if the highest token logprobs.
+	 */
 	best_of?: number;
+	/**
+	 * Whether to return decoder input token logprobs and ids.
+	 */
 	decoder_input_details?: boolean;
+	/**
+	 * Whether to return generation details.
+	 */
 	details?: boolean;
+	/**
+	 * Activate logits sampling.
+	 */
 	do_sample?: boolean;
+	/**
+	 * The parameter for frequency penalty. 1.0 means no penalty
+	 * Penalize new tokens based on their existing frequency in the text so far,
+	 * decreasing the model's likelihood to repeat the same line verbatim.
+	 */
 	frequency_penalty?: number;
 	grammar?: TextGenerationInputGrammarType;
+	/**
+	 * Maximum number of tokens to generate.
+	 */
 	max_new_tokens?: number;
+	/**
+	 * The parameter for repetition penalty. 1.0 means no penalty.
+	 * See [this paper](https://arxiv.org/pdf/1909.05858.pdf) for more details.
+	 */
 	repetition_penalty?: number;
+	/**
+	 * Whether to prepend the prompt to the generated text
+	 */
 	return_full_text?: boolean;
+	/**
+	 * Random sampling seed.
+	 */
 	seed?: number;
+	/**
+	 * Stop generating tokens if a member of `stop` is generated.
+	 */
 	stop?: string[];
+	/**
+	 * The value used to module the logits distribution.
+	 */
 	temperature?: number;
+	/**
+	 * The number of highest probability vocabulary tokens to keep for top-k-filtering.
+	 */
 	top_k?: number;
+	/**
+	 * The number of highest probability vocabulary tokens to keep for top-n-filtering.
+	 */
 	top_n_tokens?: number;
+	/**
+	 * Top-p value for nucleus sampling.
+	 */
 	top_p?: number;
+	/**
+	 * Truncate inputs tokens to the given size.
+	 */
 	truncate?: number;
+	/**
+	 * Typical Decoding mass
+	 * See [Typical Decoding for Natural Language Generation](https://arxiv.org/abs/2202.00666)
+	 * for more information.
+	 */
 	typical_p?: number;
+	/**
+	 * Watermarking with [A Watermark for Large Language
+	 * Models](https://arxiv.org/abs/2301.10226).
+	 */
 	watermark?: boolean;
 	[property: string]: unknown;
 }
@@ -125,6 +186,7 @@ export interface TextGenerationStreamOutput {
 export interface TextGenerationStreamOutputStreamDetails {
 	finish_reason: TextGenerationOutputFinishReason;
 	generated_tokens: number;
+	input_length: number;
 	seed?: number;
 	[property: string]: unknown;
 }

package/src/tasks/text-generation/spec/input.json

@@ -22,8 +22,16 @@
 		"TextGenerationInputGenerateParameters": {
 			"type": "object",
 			"properties": {
+				"adapter_id": {
+					"type": "string",
+					"description": "Lora adapter id",
+					"default": "null",
+					"example": "null",
+					"nullable": true
+				},
 				"best_of": {
 					"type": "integer",
+					"description": "Generate best_of sequences and return the one if the highest token logprobs.",
 					"default": "null",
 					"example": 1,
 					"nullable": true,
@@ -32,20 +40,24 @@
 				},
 				"decoder_input_details": {
 					"type": "boolean",
+					"description": "Whether to return decoder input token logprobs and ids.",
 					"default": "false"
 				},
 				"details": {
 					"type": "boolean",
+					"description": "Whether to return generation details.",
 					"default": "true"
 				},
 				"do_sample": {
 					"type": "boolean",
+					"description": "Activate logits sampling.",
 					"default": "false",
 					"example": true
 				},
 				"frequency_penalty": {
 					"type": "number",
 					"format": "float",
+					"description": "The parameter for frequency penalty. 1.0 means no penalty\nPenalize new tokens based on their existing frequency in the text so far,\ndecreasing the model's likelihood to repeat the same line verbatim.",
 					"default": "null",
 					"example": 0.1,
 					"nullable": true,
@@ -63,6 +75,7 @@
 				"max_new_tokens": {
 					"type": "integer",
 					"format": "int32",
+					"description": "Maximum number of tokens to generate.",
 					"default": "100",
 					"example": "20",
 					"nullable": true,
@@ -71,6 +84,7 @@
 				"repetition_penalty": {
 					"type": "number",
 					"format": "float",
+					"description": "The parameter for repetition penalty. 1.0 means no penalty.\nSee [this paper](https://arxiv.org/pdf/1909.05858.pdf) for more details.",
 					"default": "null",
 					"example": 1.03,
 					"nullable": true,
@@ -78,6 +92,7 @@
 				},
 				"return_full_text": {
 					"type": "boolean",
+					"description": "Whether to prepend the prompt to the generated text",
 					"default": "null",
 					"example": false,
 					"nullable": true
@@ -85,6 +100,7 @@
 				"seed": {
 					"type": "integer",
 					"format": "int64",
+					"description": "Random sampling seed.",
 					"default": "null",
 					"example": "null",
 					"nullable": true,
@@ -96,12 +112,14 @@
 					"items": {
 						"type": "string"
 					},
+					"description": "Stop generating tokens if a member of `stop` is generated.",
 					"example": ["photographer"],
 					"maxItems": 4
 				},
 				"temperature": {
 					"type": "number",
 					"format": "float",
+					"description": "The value used to module the logits distribution.",
 					"default": "null",
 					"example": 0.5,
 					"nullable": true,
@@ -110,6 +128,7 @@
 				"top_k": {
 					"type": "integer",
 					"format": "int32",
+					"description": "The number of highest probability vocabulary tokens to keep for top-k-filtering.",
 					"default": "null",
 					"example": 10,
 					"nullable": true,
@@ -118,6 +137,7 @@
 				"top_n_tokens": {
 					"type": "integer",
 					"format": "int32",
+					"description": "The number of highest probability vocabulary tokens to keep for top-n-filtering.",
 					"default": "null",
 					"example": 5,
 					"nullable": true,
@@ -127,6 +147,7 @@
 				"top_p": {
 					"type": "number",
 					"format": "float",
+					"description": "Top-p value for nucleus sampling.",
 					"default": "null",
 					"example": 0.95,
 					"nullable": true,
@@ -135,6 +156,7 @@
 				},
 				"truncate": {
 					"type": "integer",
+					"description": "Truncate inputs tokens to the given size.",
 					"default": "null",
 					"example": "null",
 					"nullable": true,
@@ -143,6 +165,7 @@
 				"typical_p": {
 					"type": "number",
 					"format": "float",
+					"description": "Typical Decoding mass\nSee [Typical Decoding for Natural Language Generation](https://arxiv.org/abs/2202.00666) for more information.",
 					"default": "null",
 					"example": 0.95,
 					"nullable": true,
@@ -151,6 +174,7 @@
 				},
 				"watermark": {
 					"type": "boolean",
+					"description": "Watermarking with [A Watermark for Large Language Models](https://arxiv.org/abs/2301.10226).",
 					"default": "false",
 					"example": true
 				}

package/src/tasks/text-generation/spec/stream_output.json

@@ -39,7 +39,7 @@
 	"$defs": {
 		"TextGenerationStreamOutputStreamDetails": {
 			"type": "object",
-			"required": ["finish_reason", "generated_tokens"],
+			"required": ["finish_reason", "generated_tokens", "input_length"],
 			"properties": {
 				"finish_reason": {
 					"$ref": "#/$defs/TextGenerationStreamOutputFinishReason"
@@ -50,6 +50,12 @@
 					"example": 1,
 					"minimum": 0
 				},
+				"input_length": {
+					"type": "integer",
+					"format": "int32",
+					"example": 1,
+					"minimum": 0
+				},
 				"seed": {
 					"type": "integer",
 					"format": "int64",

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

@@ -68,7 +68,7 @@ export interface GenerationParameters {
 	 */
 	max_length?: number;
 	/**
-	 * The maximum number of tokens to generate. Takes precedence over maxLength.
+	 * The maximum number of tokens to generate. Takes precedence over max_length.
 	 */
 	max_new_tokens?: number;
 	/**
@@ -76,7 +76,7 @@ export interface GenerationParameters {
 	 */
 	min_length?: number;
 	/**
-	 * The minimum number of tokens to generate. Takes precedence over maxLength.
+	 * The minimum number of tokens to generate. Takes precedence over min_length.
 	 */
 	min_new_tokens?: number;
 	/**

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

@@ -5,9 +5,7 @@
  */
 
 /**
- * Inputs for Text to Speech inference
- *
- * Inputs for Text To Audio inference
+ * Inputs for Text To Speech inference
  */
 export interface TextToSpeechInput {
 	/**
@@ -17,16 +15,16 @@ export interface TextToSpeechInput {
 	/**
 	 * Additional inference parameters
 	 */
-	parameters?: TextToAudioParameters;
+	parameters?: TextToSpeechParameters;
 	[property: string]: unknown;
 }
 
 /**
  * Additional inference parameters
  *
- * Additional inference parameters for Text To Audio
+ * Additional inference parameters for Text To Speech
  */
-export interface TextToAudioParameters {
+export interface TextToSpeechParameters {
 	/**
 	 * Parametrization of the text generation process
 	 */
@@ -70,7 +68,7 @@ export interface GenerationParameters {
 	 */
 	max_length?: number;
 	/**
-	 * The maximum number of tokens to generate. Takes precedence over maxLength.
+	 * The maximum number of tokens to generate. Takes precedence over max_length.
 	 */
 	max_new_tokens?: number;
 	/**
@@ -78,7 +76,7 @@ export interface GenerationParameters {
 	 */
 	min_length?: number;
 	/**
-	 * The minimum number of tokens to generate. Takes precedence over maxLength.
+	 * The minimum number of tokens to generate. Takes precedence over min_length.
 	 */
 	min_new_tokens?: number;
 	/**

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

@@ -1,7 +1,31 @@
 {
-	"$ref": "/inference/schemas/text-to-audio/input.json",
 	"$id": "/inference/schemas/text-to-speech/input.json",
 	"$schema": "http://json-schema.org/draft-06/schema#",
+	"description": "Inputs for Text To Speech inference",
 	"title": "TextToSpeechInput",
-	"description": "Inputs for Text to Speech inference"
+	"type": "object",
+	"properties": {
+		"inputs": {
+			"description": "The input text data",
+			"type": "string"
+		},
+		"parameters": {
+			"description": "Additional inference parameters",
+			"$ref": "#/$defs/TextToSpeechParameters"
+		}
+	},
+	"$defs": {
+		"TextToSpeechParameters": {
+			"title": "TextToSpeechParameters",
+			"description": "Additional inference parameters for Text To Speech",
+			"type": "object",
+			"properties": {
+				"generate": {
+					"description": "Parametrization of the text generation process",
+					"$ref": "/inference/schemas/common-definitions.json#/definitions/GenerationParameters"
+				}
+			}
+		}
+	},
+	"required": ["inputs"]
 }