Merge branch 'main' into cerebras-provider

Wauplin · web-flow · commit 93093edc644b · 2025-03-04T10:32:47.000+01:00
diff --git a/packages/ollama-utils/package.json b/packages/ollama-utils/package.json
@@ -1,7 +1,7 @@
 {
 	"name": "@huggingface/ollama-utils",
 	"packageManager": "pnpm@8.10.5",
-	"version": "0.0.2",
+	"version": "0.0.3",
 	"description": "Various utilities for maintaining Ollama compatibility with models on Hugging Face hub",
 	"repository": "https://github.com/huggingface/huggingface.js.git",
 	"publishConfig": {
diff --git a/packages/ollama-utils/src/chat-template-automap.ts b/packages/ollama-utils/src/chat-template-automap.ts
diff --git a/packages/tasks-gen/scripts/generate-snippets-fixtures.ts b/packages/tasks-gen/scripts/generate-snippets-fixtures.ts
@@ -90,6 +90,17 @@ const TEST_CASES: {
 		providers: ["hf-inference", "fal-ai"],
 		languages: ["sh", "js", "py"],
 	},
+	{
+		testName: "text-to-video",
+		model: {
+			id: "tencent/HunyuanVideo",
+			pipeline_tag: "text-to-video",
+			tags: [],
+			inference: "",
+		},
+		providers: ["replicate", "fal-ai"],
+		languages: ["js", "py"],
+	},
 	{
 		testName: "text-classification",
 		model: {
diff --git a/packages/tasks-gen/snippets-fixtures/text-to-video/0.huggingface.js.fal-ai.js b/packages/tasks-gen/snippets-fixtures/text-to-video/0.huggingface.js.fal-ai.js
@@ -0,0 +1,11 @@
+import { HfInference } from "@huggingface/inference";
+
+const client = new HfInference("api_token");
+
+const video = await client.textToVideo({
+	model: "tencent/HunyuanVideo",
+	provider: "fal-ai",
+	inputs: "A young man walking on the street",
+	parameters: { num_inference_steps: 5 },
+});
+// Use the generated video (it's a Blob)
diff --git a/packages/tasks-gen/snippets-fixtures/text-to-video/0.huggingface.js.replicate.js b/packages/tasks-gen/snippets-fixtures/text-to-video/0.huggingface.js.replicate.js
@@ -0,0 +1,11 @@
+import { HfInference } from "@huggingface/inference";
+
+const client = new HfInference("api_token");
+
+const video = await client.textToVideo({
+	model: "tencent/HunyuanVideo",
+	provider: "replicate",
+	inputs: "A young man walking on the street",
+	parameters: { num_inference_steps: 5 },
+});
+// Use the generated video (it's a Blob)
diff --git a/packages/tasks-gen/snippets-fixtures/text-to-video/0.huggingface_hub.fal-ai.py b/packages/tasks-gen/snippets-fixtures/text-to-video/0.huggingface_hub.fal-ai.py
@@ -0,0 +1,11 @@
+from huggingface_hub import InferenceClient
+
+client = InferenceClient(
+	provider="fal-ai",
+	api_key="api_token"
+)
+
+video = client.text_to_video(
+	"A young man walking on the street",
+	model="tencent/HunyuanVideo"
+)
diff --git a/packages/tasks-gen/snippets-fixtures/text-to-video/0.huggingface_hub.replicate.py b/packages/tasks-gen/snippets-fixtures/text-to-video/0.huggingface_hub.replicate.py
@@ -0,0 +1,11 @@
+from huggingface_hub import InferenceClient
+
+client = InferenceClient(
+	provider="replicate",
+	api_key="api_token"
+)
+
+video = client.text_to_video(
+	"A young man walking on the street",
+	model="tencent/HunyuanVideo"
+)
diff --git a/packages/tasks/package.json b/packages/tasks/package.json
@@ -1,7 +1,7 @@
 {
 	"name": "@huggingface/tasks",
 	"packageManager": "pnpm@8.10.5",
-	"version": "0.16.5",
+	"version": "0.16.6",
 	"description": "List of ML tasks for huggingface.co/tasks",
 	"repository": "https://github.com/huggingface/huggingface.js.git",
 	"publishConfig": {
diff --git a/packages/tasks/src/model-libraries.ts b/packages/tasks/src/model-libraries.ts
@@ -394,6 +394,12 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
 		repoUrl: "https://github.com/Tencent/HunyuanDiT",
 		countDownloads: `path:"pytorch_model_ema.pt" OR path:"pytorch_model_distill.pt"`,
 	},
+	"hunyuan3d-2": {
+		prettyLabel: "Hunyuan3D-2",
+		repoName: "Hunyuan3D-2",
+		repoUrl: "https://github.com/Tencent/Hunyuan3D-2",
+		countDownloads: `path:"model_index.json" OR path:"config.yaml"`,
+	},
 	imstoucan: {
 		prettyLabel: "IMS Toucan",
 		repoName: "IMS-Toucan",
diff --git a/packages/tasks/src/snippets/inputs.ts b/packages/tasks/src/snippets/inputs.ts
@@ -96,6 +96,8 @@ const inputsAudioClassification = () => `"sample1.flac"`;
 
 const inputsTextToImage = () => `"Astronaut riding a horse"`;
 
+const inputsTextToVideo = () => `"A young man walking on the street"`;
+
 const inputsTextToSpeech = () => `"The answer to the universe is 42"`;
 
 const inputsTextToAudio = () => `"liquid drum and bass, atmospheric synths, airy sounds"`;
@@ -130,6 +132,7 @@ const modelInputSnippets: {
 	"text-generation": inputsTextGeneration,
 	"image-text-to-text": inputsTextGeneration,
 	"text-to-image": inputsTextToImage,
+	"text-to-video": inputsTextToVideo,
 	"text-to-speech": inputsTextToSpeech,
 	"text-to-audio": inputsTextToAudio,
 	"text2text-generation": inputsText2TextGeneration,
diff --git a/packages/tasks/src/snippets/js.ts b/packages/tasks/src/snippets/js.ts
@@ -275,6 +275,33 @@ query({"inputs": ${getModelInputSnippet(model)}}).then((response) => {
 	];
 };
 
+export const snippetTextToVideo = (
+	model: ModelDataMinimal,
+	accessToken: string,
+	provider: SnippetInferenceProvider
+): InferenceSnippet[] => {
+	return ["fal-ai", "replicate"].includes(provider)
+		? [
+				{
+					client: "huggingface.js",
+					content: `\
+import { HfInference } from "@huggingface/inference";
+
+const client = new HfInference("${accessToken || `{API_TOKEN}`}");
+
+const video = await client.textToVideo({
+	model: "${model.id}",
+	provider: "${provider}",
+	inputs: ${getModelInputSnippet(model)},
+	parameters: { num_inference_steps: 5 },
+});
+// Use the generated video (it's a Blob)
+`,
+				},
+		  ]
+		: [];
+};
+
 export const snippetTextToAudio = (
 	model: ModelDataMinimal,
 	accessToken: string,
@@ -420,6 +447,7 @@ export const jsSnippets: Partial<
 	"sentence-similarity": snippetBasic,
 	"automatic-speech-recognition": snippetAutomaticSpeechRecognition,
 	"text-to-image": snippetTextToImage,
+	"text-to-video": snippetTextToVideo,
 	"text-to-speech": snippetTextToAudio,
 	"text-to-audio": snippetTextToAudio,
 	"audio-to-audio": snippetFile,
diff --git a/packages/tasks/src/snippets/python.ts b/packages/tasks/src/snippets/python.ts
@@ -308,6 +308,27 @@ image = Image.open(io.BytesIO(image_bytes))`,
 	];
 };
 
+export const snippetTextToVideo = (
+	model: ModelDataMinimal,
+	accessToken: string,
+	provider: SnippetInferenceProvider
+): InferenceSnippet[] => {
+	return ["fal-ai", "replicate"].includes(provider)
+		? [
+				{
+					client: "huggingface_hub",
+					content: `\
+${snippetImportInferenceClient(accessToken, provider)}
+
+video = client.text_to_video(
+	${getModelInputSnippet(model)},
+	model="${model.id}"
+)`,
+				},
+		  ]
+		: [];
+};
+
 export const snippetTabular = (model: ModelDataMinimal): InferenceSnippet[] => {
 	return [
 		{
@@ -412,6 +433,7 @@ export const pythonSnippets: Partial<
 	"sentence-similarity": snippetBasic,
 	"automatic-speech-recognition": snippetFile,
 	"text-to-image": snippetTextToImage,
+	"text-to-video": snippetTextToVideo,
 	"text-to-speech": snippetTextToAudio,
 	"text-to-audio": snippetTextToAudio,
 	"audio-to-audio": snippetFile,
diff --git a/packages/tasks/src/tasks/chat-completion/inference.ts b/packages/tasks/src/tasks/chat-completion/inference.ts
@@ -105,9 +105,10 @@ export interface ChatCompletionInput {
 	[property: string]: unknown;
 }
 export interface ChatCompletionInputMessage {
-	content: ChatCompletionInputMessageContent;
+	content?: ChatCompletionInputMessageContent;
 	name?: string;
 	role: string;
+	tool_calls?: ChatCompletionInputToolCall[];
 	[property: string]: unknown;
 }
 export type ChatCompletionInputMessageContent = ChatCompletionInputMessageChunk[] | string;
@@ -122,6 +123,18 @@ export interface ChatCompletionInputURL {
 	[property: string]: unknown;
 }
 export type ChatCompletionInputMessageChunkType = "text" | "image_url";
+export interface ChatCompletionInputToolCall {
+	function: ChatCompletionInputFunctionDefinition;
+	id: string;
+	type: string;
+	[property: string]: unknown;
+}
+export interface ChatCompletionInputFunctionDefinition {
+	arguments: unknown;
+	description?: string;
+	name: string;
+	[property: string]: unknown;
+}
 export interface ChatCompletionInputGrammarType {
 	type: ChatCompletionInputGrammarTypeType;
 	/**
@@ -170,12 +183,6 @@ export interface ChatCompletionInputTool {
 	type: string;
 	[property: string]: unknown;
 }
-export interface ChatCompletionInputFunctionDefinition {
-	arguments: unknown;
-	description?: string;
-	name: string;
-	[property: string]: unknown;
-}
 /**
  * Chat Completion Output.
  *
@@ -217,6 +224,7 @@ export interface ChatCompletionOutputTopLogprob {
 export interface ChatCompletionOutputMessage {
 	content?: string;
 	role: string;
+	tool_call_id?: string;
 	tool_calls?: ChatCompletionOutputToolCall[];
 	[property: string]: unknown;
 }
@@ -264,6 +272,7 @@ export interface ChatCompletionStreamOutputChoice {
 export interface ChatCompletionStreamOutputDelta {
 	content?: string;
 	role: string;
+	tool_call_id?: string;
 	tool_calls?: ChatCompletionStreamOutputDeltaToolCall;
 	[property: string]: unknown;
 }
diff --git a/packages/tasks/src/tasks/chat-completion/spec/input.json b/packages/tasks/src/tasks/chat-completion/spec/input.json
@@ -151,24 +151,54 @@
 	},
 	"$defs": {
 		"ChatCompletionInputMessage": {
-			"type": "object",
-			"required": ["role", "content"],
-			"properties": {
-				"content": {
-					"$ref": "#/$defs/ChatCompletionInputMessageContent"
-				},
-				"name": {
-					"type": "string",
-					"example": "\"David\"",
-					"nullable": true
+			"allOf": [
+				{
+					"$ref": "#/$defs/ChatCompletionInputMessageBody"
 				},
-				"role": {
-					"type": "string",
-					"example": "user"
+				{
+					"type": "object",
+					"required": ["role"],
+					"properties": {
+						"name": {
+							"type": "string",
+							"example": "\"David\"",
+							"nullable": true
+						},
+						"role": {
+							"type": "string",
+							"example": "user"
+						}
+					}
 				}
-			},
+			],
 			"title": "ChatCompletionInputMessage"
 		},
+		"ChatCompletionInputMessageBody": {
+			"oneOf": [
+				{
+					"type": "object",
+					"required": ["content"],
+					"properties": {
+						"content": {
+							"$ref": "#/$defs/ChatCompletionInputMessageContent"
+						}
+					}
+				},
+				{
+					"type": "object",
+					"required": ["tool_calls"],
+					"properties": {
+						"tool_calls": {
+							"type": "array",
+							"items": {
+								"$ref": "#/$defs/ChatCompletionInputToolCall"
+							}
+						}
+					}
+				}
+			],
+			"title": "ChatCompletionInputMessageBody"
+		},
 		"ChatCompletionInputMessageContent": {
 			"oneOf": [
 				{
@@ -227,6 +257,37 @@
 			},
 			"title": "ChatCompletionInputUrl"
 		},
+		"ChatCompletionInputToolCall": {
+			"type": "object",
+			"required": ["id", "type", "function"],
+			"properties": {
+				"function": {
+					"$ref": "#/$defs/ChatCompletionInputFunctionDefinition"
+				},
+				"id": {
+					"type": "string"
+				},
+				"type": {
+					"type": "string"
+				}
+			},
+			"title": "ChatCompletionInputToolCall"
+		},
+		"ChatCompletionInputFunctionDefinition": {
+			"type": "object",
+			"required": ["name", "arguments"],
+			"properties": {
+				"arguments": {},
+				"description": {
+					"type": "string",
+					"nullable": true
+				},
+				"name": {
+					"type": "string"
+				}
+			},
+			"title": "ChatCompletionInputFunctionDefinition"
+		},
 		"ChatCompletionInputGrammarType": {
 			"oneOf": [
 				{
@@ -326,21 +387,6 @@
 				}
 			},
 			"title": "ChatCompletionInputTool"
-		},
-		"ChatCompletionInputFunctionDefinition": {
-			"type": "object",
-			"required": ["name", "arguments"],
-			"properties": {
-				"arguments": {},
-				"description": {
-					"type": "string",
-					"nullable": true
-				},
-				"name": {
-					"type": "string"
-				}
-			},
-			"title": "ChatCompletionInputFunctionDefinition"
 		}
 	}
 }
diff --git a/packages/tasks/src/tasks/chat-completion/spec/output.json b/packages/tasks/src/tasks/chat-completion/spec/output.json
@@ -128,6 +128,10 @@
 				"role": {
 					"type": "string",
 					"example": "user"
+				},
+				"tool_call_id": {
+					"type": "string",
+					"nullable": true
 				}
 			},
 			"title": "ChatCompletionOutputTextMessage"
diff --git a/packages/tasks/src/tasks/chat-completion/spec/stream_output.json b/packages/tasks/src/tasks/chat-completion/spec/stream_output.json
diff --git a/packages/tasks/src/tasks/text-to-video/data.ts b/packages/tasks/src/tasks/text-to-video/data.ts

Original file line number	Diff line number	Diff line change
`@@ -1,7 +1,7 @@`
`1`	`1`	`{`
`2`	`2`	`"name": "@huggingface/ollama-utils",`
`3`	`3`	`"packageManager": "[email protected]",`
`4`		`- "version": "0.0.2",`
	`4`	`+ "version": "0.0.3",`
`5`	`5`	`"description": "Various utilities for maintaining Ollama compatibility with models on Hugging Face hub",`
`6`	`6`	`"repository": "https://github.com/huggingface/huggingface.js.git",`
`7`	`7`	`"publishConfig": {`
Original file line number	Diff line number	Diff line change
`@@ -1,7 +1,7 @@`
`1`	`1`	`{`
`2`	`2`	`"name": "@huggingface/tasks",`
`3`	`3`	`"packageManager": "[email protected]",`
`4`		`- "version": "0.16.5",`
	`4`	`+ "version": "0.16.6",`
`5`	`5`	`"description": "List of ML tasks for huggingface.co/tasks",`
`6`	`6`	`"repository": "https://github.com/huggingface/huggingface.js.git",`
`7`	`7`	`"publishConfig": {`