Merge branch 'main' into feat/support-novita-async-t2v

Author: viktor2077

.github/workflows/tiny-agents-publish.yml

@@ -0,0 +1,77 @@
+name: Tiny Agents - Version and Release
+
+on:
+  workflow_dispatch:
+    inputs:
+      newversion:
+        type: choice
+        description: "Semantic Version Bump Type"
+        default: patch
+        options:
+          - patch
+          - minor
+          - major
+      bypass_deps_check:
+        type: boolean
+        description: "Bypass dependency checking"
+        default: false
+
+concurrency:
+  group: "push-to-main" # Consider changing this if tiny-agents has its own release concurrency group
+
+defaults:
+  run:
+    working-directory: packages/tiny-agents
+
+jobs:
+  version_and_release:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          token: ${{ secrets.BOT_ACCESS_TOKEN }}
+      - run: npm install -g corepack@latest && corepack enable
+      - uses: actions/setup-node@v3
+        with:
+          node-version: "20"
+          cache: "pnpm"
+          cache-dependency-path: |
+            packages/tiny-agents/pnpm-lock.yaml
+            packages/doc-internal/pnpm-lock.yaml
+          registry-url: "https://registry.npmjs.org"
+      - run: pnpm install
+      - run: git config --global user.name machineuser
+      - run: git config --global user.email [email protected]
+      - run: |
+          PACKAGE_VERSION=$(node -p "require('./package.json').version")
+          BUMPED_VERSION=$(node -p "require('semver').inc('$PACKAGE_VERSION', '${{ github.event.inputs.newversion }}')")
+          # Update package.json with the new version
+          node -e "const fs = require('fs'); const package = JSON.parse(fs.readFileSync('./package.json')); package.version = '$BUMPED_VERSION'; fs.writeFileSync('./package.json', JSON.stringify(package, null, '\t') + '\n');"
+          pnpm --filter doc-internal run fix-cdn-versions
+          git add ../..
+          git commit -m "🔖 @huggingface/tiny-agents $BUMPED_VERSION"
+          git tag "tiny-agents-v$BUMPED_VERSION"
+
+      # Add checks for dependencies if needed, similar to hub-publish.yml
+      - if: ${{ !github.event.inputs.bypass_deps_check }}
+        name: "Check Deps are published before publishing this package"
+        run: pnpm -w check-deps inference && pnpm -w check-deps tasks # Review if these specific deps apply to tiny-agents
+
+      - run: pnpm publish --no-git-checks .
+        env:
+          NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}
+      - run: (git pull --rebase && git push --follow-tags) || (git pull --rebase && git push --follow-tags)
+      # hack - reuse actions/setup-node@v3 just to set a new registry
+      - uses: actions/setup-node@v3
+        with:
+          node-version: "20"
+          registry-url: "https://npm.pkg.github.com"
+      # Disable for now, until github supports PATs for writing github packages (https://github.com/github/roadmap/issues/558)
+      # - run: pnpm publish --no-git-checks .
+      #   env:
+      #     NODE_AUTH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      - name: "Update Doc"
+        uses: peter-evans/repository-dispatch@v2
+        with:
+          event-type: doc-build
+          token: ${{ secrets.BOT_ACCESS_TOKEN }}

README.md

@@ -97,7 +97,7 @@ You can run our packages with vanilla JS, without any bundler, by using a CDN or
 
 ```html
 <script type="module">
-    import { InferenceClient } from 'https://cdn.jsdelivr.net/npm/@huggingface/[email protected].1/+esm';
+    import { InferenceClient } from 'https://cdn.jsdelivr.net/npm/@huggingface/[email protected].2/+esm';
     import { createRepo, commit, deleteRepo, listFiles } from "https://cdn.jsdelivr.net/npm/@huggingface/[email protected]/+esm";
 </script>
 ```
@@ -241,8 +241,8 @@ const agent = new Agent({
   ],
 });
 
-
 await agent.loadTools();
+
 for await (const chunk of agent.run("What are the top 5 trending models on Hugging Face?")) {
     if ("choices" in chunk) {
         const delta = chunk.choices[0]?.delta;

packages/inference/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@huggingface/inference",
-	"version": "3.13.1",
+	"version": "3.13.2",
 	"packageManager": "[email protected]",
 	"license": "MIT",
 	"author": "Hugging Face and Tim Mikeladze <[email protected]>",

packages/inference/src/lib/getProviderHelper.ts

@@ -47,7 +47,7 @@ import type {
 import * as Replicate from "../providers/replicate";
 import * as Sambanova from "../providers/sambanova";
 import * as Together from "../providers/together";
-import type { InferenceProvider, InferenceTask } from "../types";
+import type { InferenceProvider, InferenceProviderOrPolicy, InferenceTask } from "../types";
 
 export const PROVIDERS: Record<InferenceProvider, Partial<Record<InferenceTask, TaskProviderHelper>>> = {
 	"black-forest-labs": {
@@ -153,128 +153,132 @@ export const PROVIDERS: Record<InferenceProvider, Partial<Record<InferenceTask,
  * Get provider helper instance by name and task
  */
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "text-to-image"
 ): TextToImageTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "conversational"
 ): ConversationalTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "text-generation"
 ): TextGenerationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "text-to-speech"
 ): TextToSpeechTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "text-to-audio"
 ): TextToAudioTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "automatic-speech-recognition"
 ): AutomaticSpeechRecognitionTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "text-to-video"
 ): TextToVideoTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "text-classification"
 ): TextClassificationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "question-answering"
 ): QuestionAnsweringTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "audio-classification"
 ): AudioClassificationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "audio-to-audio"
 ): AudioToAudioTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "fill-mask"
 ): FillMaskTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "feature-extraction"
 ): FeatureExtractionTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "image-classification"
 ): ImageClassificationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "image-segmentation"
 ): ImageSegmentationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "document-question-answering"
 ): DocumentQuestionAnsweringTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "image-to-text"
 ): ImageToTextTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "object-detection"
 ): ObjectDetectionTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "zero-shot-image-classification"
 ): ZeroShotImageClassificationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "zero-shot-classification"
 ): ZeroShotClassificationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "image-to-image"
 ): ImageToImageTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "sentence-similarity"
 ): SentenceSimilarityTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "table-question-answering"
 ): TableQuestionAnsweringTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "tabular-classification"
 ): TabularClassificationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "tabular-regression"
 ): TabularRegressionTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "token-classification"
 ): TokenClassificationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "translation"
 ): TranslationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "summarization"
 ): SummarizationTaskHelper & TaskProviderHelper;
 export function getProviderHelper(
-	provider: InferenceProvider,
+	provider: InferenceProviderOrPolicy,
 	task: "visual-question-answering"
 ): VisualQuestionAnsweringTaskHelper & TaskProviderHelper;
-export function getProviderHelper(provider: InferenceProvider, task: InferenceTask | undefined): TaskProviderHelper;
+export function getProviderHelper(
+	provider: InferenceProviderOrPolicy,
+	task: InferenceTask | undefined
+): TaskProviderHelper;
 
-export function getProviderHelper(provider: InferenceProvider, task: InferenceTask | undefined): TaskProviderHelper {
-	if (provider === "hf-inference") {
-		if (!task) {
-			return new HFInference.HFInferenceTask();
-		}
+export function getProviderHelper(
+	provider: InferenceProviderOrPolicy,
+	task: InferenceTask | undefined
+): TaskProviderHelper {
+	if ((provider === "hf-inference" && !task) || provider === "auto") {
+		return new HFInference.HFInferenceTask();
 	}
 	if (!task) {
 		throw new Error("you need to provide a task name when using an external provider, e.g. 'text-to-image'");

packages/inference/src/snippets/getInferenceSnippets.ts

@@ -144,8 +144,7 @@ const snippetGenerator = (templateName: string, inputPreparationFn?: InputPrepar
 		}
 		let providerHelper: ReturnType<typeof getProviderHelper>;
 		try {
-			/// For the "auto" provider policy we use hf-inference snippets
-			providerHelper = getProviderHelper(provider === "auto" ? "hf-inference" : provider, task);
+			providerHelper = getProviderHelper(provider, task);
 		} catch (e) {
 			console.error(`Failed to get provider helper for ${provider} (${task})`, e);
 			return [];

packages/mcp-client/src/index.ts

@@ -1,3 +1,4 @@
 export * from "./McpClient";
 export * from "./Agent";
 export type { ChatCompletionInputMessageTool } from "./McpClient";
+export type { ServerConfig } from "./types";

packages/tasks/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "@huggingface/tasks",
 	"packageManager": "[email protected]",
-	"version": "0.19.6",
+	"version": "0.19.7",
 	"description": "List of ML tasks for huggingface.co/tasks",
 	"repository": "https://github.com/huggingface/huggingface.js.git",
 	"publishConfig": {

packages/tasks/src/model-libraries-snippets.ts

@@ -593,6 +593,39 @@ export const keras_hub = (model: ModelData): string[] => {
 	return snippets;
 };
 
+export const kimi_audio = (model: ModelData): string[] => [
+	`# Example usage for KimiAudio
+# pip install git+https://github.com/MoonshotAI/Kimi-Audio.git
+
+from kimia_infer.api.kimia import KimiAudio
+
+model = KimiAudio(model_path="${model.id}", load_detokenizer=True)
+
+sampling_params = {
+    "audio_temperature": 0.8,
+    "audio_top_k": 10,
+    "text_temperature": 0.0,
+    "text_top_k": 5,
+}
+
+# For ASR
+asr_audio = "asr_example.wav"
+messages_asr = [
+    {"role": "user", "message_type": "text", "content": "Please transcribe the following audio:"},
+    {"role": "user", "message_type": "audio", "content": asr_audio}
+]
+_, text = model.generate(messages_asr, **sampling_params, output_type="text")
+print(text)
+
+# For Q&A
+qa_audio = "qa_example.wav"
+messages_conv = [{"role": "user", "message_type": "audio", "content": qa_audio}]
+wav, text = model.generate(messages_conv, **sampling_params, output_type="both")
+sf.write("output_audio.wav", wav.cpu().view(-1).numpy(), 24000)
+print(text)
+`,
+];
+
 export const lightning_ir = (model: ModelData): string[] => {
 	if (model.tags.includes("bi-encoder")) {
 		return [
@@ -1164,10 +1197,11 @@ export const transformers = (model: ModelData): string[] => {
 					].join("\n"),
 					"]"
 				);
+				pipelineSnippet.push("pipe(text=messages)");
 			} else {
 				pipelineSnippet.push("messages = [", '    {"role": "user", "content": "Who are you?"},', "]");
+				pipelineSnippet.push("pipe(messages)");
 			}
-			pipelineSnippet.push("pipe(messages)");
 		}
 
 		return [pipelineSnippet.join("\n"), autoSnippet];
@@ -1354,15 +1388,14 @@ model = SwarmFormerModel.from_pretrained("${model.id}")
 
 const mlx_unknown = (model: ModelData): string[] => [
 	`# Download the model from the Hub
-pip install huggingface_hub hf_transfer
+pip install huggingface_hub[hf_xet]
 
-export HF_HUB_ENABLE_HF_TRANSFER=1
 huggingface-cli download --local-dir ${nameWithoutNamespace(model.id)} ${model.id}`,
 ];
 
 const mlxlm = (model: ModelData): string[] => [
 	`# Make sure mlx-lm is installed
-pip install --upgrade mlx-lm
+# pip install --upgrade mlx-lm
 
 # Generate text with mlx-lm
 from mlx_lm import load, generate
@@ -1375,7 +1408,7 @@ text = generate(model, tokenizer, prompt=prompt, verbose=True)`,
 
 const mlxchat = (model: ModelData): string[] => [
 	`# Make sure mlx-lm is installed
-pip install --upgrade mlx-lm
+# pip install --upgrade mlx-lm
 
 # Generate text with mlx-lm
 from mlx_lm import load, generate
@@ -1392,7 +1425,9 @@ text = generate(model, tokenizer, prompt=prompt, verbose=True)`,
 ];
 
 const mlxvlm = (model: ModelData): string[] => [
-	`Make sure mlx-vlm is installed
+	`# Make sure mlx-vlm is installed
+# pip install --upgrade mlx-vlm
+
 from mlx_vlm import load, generate
 from mlx_vlm.prompt_utils import apply_chat_template
 from mlx_vlm.utils import load_config