Merge branch 'main' into patch-1

huggingface · Aug 23, 2024 · af32ffa · af32ffa
2 parents 74891b9 + 6e47ae8
commit af32ffa
Show file tree

Hide file tree

Showing 5 changed files with 38 additions and 17 deletions.
diff --git a/packages/tasks/package.json b/packages/tasks/package.json
@@ -1,7 +1,7 @@
 {
 	"name": "@huggingface/tasks",
 	"packageManager": "[email protected]",
-	"version": "0.11.10",
+	"version": "0.11.11",
 	"description": "List of ML tasks for huggingface.co/tasks",
 	"repository": "https://github.com/huggingface/huggingface.js.git",
 	"publishConfig": {

diff --git a/packages/tasks/src/dataset-libraries.ts b/packages/tasks/src/dataset-libraries.ts
@@ -77,6 +77,12 @@ export const DATASET_LIBRARIES_UI_ELEMENTS = {
 		repoUrl: "https://github.com/pola-rs/polars",
 		docsUrl: "https://huggingface.co/docs/hub/datasets-polars",
 	},
+	duckdb: {
+		prettyLabel: "DuckDB",
+		repoName: "duckdb",
+		repoUrl: "https://github.com/duckdb/duckdb",
+		docsUrl: "https://huggingface.co/docs/hub/datasets-duckdb",
+	},
 } satisfies Record<string, DatasetLibraryUiElement>;
 
 /// List of the dataset libraries supported by the Hub

diff --git a/packages/tasks/src/local-apps.ts b/packages/tasks/src/local-apps.ts
@@ -58,10 +58,15 @@ export type LocalApp = {
 	  }
 );
 
+// eslint-disable-next-line @typescript-eslint/no-unused-vars
 function isGgufModel(model: ModelData) {
 	return model.tags.includes("gguf");
 }
 
+function isLlamaCppGgufModel(model: ModelData) {
+	return !!model.gguf?.context_length;
+}
+
 const snippetLlamacpp = (model: ModelData, filepath?: string): LocalAppSnippet[] => {
 	const command = (binary: string) =>
 		[
@@ -138,56 +143,56 @@ export const LOCAL_APPS = {
 		prettyLabel: "llama.cpp",
 		docsUrl: "https://github.com/ggerganov/llama.cpp",
 		mainTask: "text-generation",
-		displayOnModelPage: isGgufModel,
+		displayOnModelPage: isLlamaCppGgufModel,
 		snippet: snippetLlamacpp,
 	},
 	lmstudio: {
 		prettyLabel: "LM Studio",
 		docsUrl: "https://lmstudio.ai",
 		mainTask: "text-generation",
-		displayOnModelPage: isGgufModel,
+		displayOnModelPage: isLlamaCppGgufModel,
 		deeplink: (model, filepath) =>
 			new URL(`lmstudio://open_from_hf?model=${model.id}${filepath ? `&file=${filepath}` : ""}`),
 	},
 	localai: {
 		prettyLabel: "LocalAI",
 		docsUrl: "https://github.com/mudler/LocalAI",
 		mainTask: "text-generation",
-		displayOnModelPage: isGgufModel,
+		displayOnModelPage: isLlamaCppGgufModel,
 		snippet: snippetLocalAI,
 	},
 	jan: {
 		prettyLabel: "Jan",
 		docsUrl: "https://jan.ai",
 		mainTask: "text-generation",
-		displayOnModelPage: isGgufModel,
+		displayOnModelPage: isLlamaCppGgufModel,
 		deeplink: (model) => new URL(`jan://models/huggingface/${model.id}`),
 	},
 	backyard: {
 		prettyLabel: "Backyard AI",
 		docsUrl: "https://backyard.ai",
 		mainTask: "text-generation",
-		displayOnModelPage: isGgufModel,
+		displayOnModelPage: isLlamaCppGgufModel,
 		deeplink: (model) => new URL(`https://backyard.ai/hf/model/${model.id}`),
 	},
 	sanctum: {
 		prettyLabel: "Sanctum",
 		docsUrl: "https://sanctum.ai",
 		mainTask: "text-generation",
-		displayOnModelPage: isGgufModel,
+		displayOnModelPage: isLlamaCppGgufModel,
 		deeplink: (model) => new URL(`sanctum://open_from_hf?model=${model.id}`),
 	},
 	jellybox: {
 		prettyLabel: "Jellybox",
 		docsUrl: "https://jellybox.com",
 		mainTask: "text-generation",
 		displayOnModelPage: (model) =>
-			isGgufModel(model) ||
+			isLlamaCppGgufModel(model) ||
 			(model.library_name === "diffusers" &&
 				model.tags.includes("safetensors") &&
 				(model.pipeline_tag === "text-to-image" || model.tags.includes("lora"))),
 		deeplink: (model) => {
-			if (isGgufModel(model)) {
+			if (isLlamaCppGgufModel(model)) {
 				return new URL(`jellybox://llm/models/huggingface/LLM/${model.id}`);
 			} else if (model.tags.includes("lora")) {
 				return new URL(`jellybox://image/models/huggingface/ImageLora/${model.id}`);
@@ -200,15 +205,15 @@ export const LOCAL_APPS = {
 		prettyLabel: "Msty",
 		docsUrl: "https://msty.app",
 		mainTask: "text-generation",
-		displayOnModelPage: isGgufModel,
+		displayOnModelPage: isLlamaCppGgufModel,
 		deeplink: (model) => new URL(`msty://models/search/hf/${model.id}`),
 	},
 	recursechat: {
 		prettyLabel: "RecurseChat",
 		docsUrl: "https://recurse.chat",
 		mainTask: "text-generation",
 		macOSOnly: true,
-		displayOnModelPage: isGgufModel,
+		displayOnModelPage: isLlamaCppGgufModel,
 		deeplink: (model) => new URL(`recursechat://new-hf-gguf-model?hf-model-id=${model.id}`),
 	},
 	drawthings: {

diff --git a/packages/tasks/src/model-data.ts b/packages/tasks/src/model-data.ts
@@ -109,6 +109,16 @@ export interface ModelData {
 	 * Example: transformers, SpeechBrain, Stanza, etc.
 	 */
 	library_name?: string;
+	safetensors?: {
+		parameters: Record<string, number>;
+		total: number;
+		sharded: boolean;
+	};
+	gguf?: {
+		total: number;
+		architecture?: string;
+		context_length?: number;
+	};
 }
 
 /**

diff --git a/packages/tasks/src/model-libraries-snippets.ts b/packages/tasks/src/model-libraries-snippets.ts
@@ -270,12 +270,12 @@ llm = Llama.from_pretrained(
 )
 
 llm.create_chat_completion(
-		messages = [
-			{
-				"role": "user",
-				"content": "What is the capital of France?"
-			}
-		]
+	messages = [
+		{
+			"role": "user",
+			"content": "What is the capital of France?"
+		}
+	]
 )`,
 ];