Finally rename HfInference to InferenceClient (#1258)

There's no reason the API should be ≠ in Python and JS

---------

Co-authored-by: vb <[email protected]>

Author: julien-c

README.md

@@ -31,6 +31,7 @@ await uploadFile({
 
 await inference.chatCompletion({
   model: "meta-llama/Llama-3.1-8B-Instruct",
+  provider: "sambanova", // or together, fal-ai, replicate, cohere …
   messages: [
     {
       role: "user",
@@ -39,11 +40,11 @@ await inference.chatCompletion({
   ],
   max_tokens: 512,
   temperature: 0.5,
-  provider: "sambanova", // or together, fal-ai, replicate, cohere …
 });
 
 await inference.textToImage({
   model: "black-forest-labs/FLUX.1-dev",
+  provider: "replicate",
   inputs: "a picture of a green bird",
 });
 
@@ -54,7 +55,7 @@ await inference.textToImage({
 
 This is a collection of JS libraries to interact with the Hugging Face API, with TS types included.
 
-- [@huggingface/inference](packages/inference/README.md): Use HF Inference API (serverless), Inference Endpoints (dedicated) and third-party Inference Providers to make calls to 100,000+ Machine Learning models
+- [@huggingface/inference](packages/inference/README.md): Use HF Inference API (serverless), Inference Endpoints (dedicated) and all supported Inference Providers to make calls to 100,000+ Machine Learning models
 - [@huggingface/hub](packages/hub/README.md): Interact with huggingface.co to create or delete repos and commit / download files
 - [@huggingface/agents](packages/agents/README.md): Interact with HF models through a natural language interface
 - [@huggingface/gguf](packages/gguf/README.md): A GGUF parser that works on remotely hosted files.
@@ -84,7 +85,7 @@ npm install @huggingface/agents
 Then import the libraries in your code:
 
 ```ts
-import { HfInference } from "@huggingface/inference";
+import { InferenceClient } from "@huggingface/inference";
 import { HfAgent } from "@huggingface/agents";
 import { createRepo, commit, deleteRepo, listFiles } from "@huggingface/hub";
 import type { RepoId } from "@huggingface/hub";
@@ -96,7 +97,7 @@ You can run our packages with vanilla JS, without any bundler, by using a CDN or
 
 ```html
 <script type="module">
-    import { HfInference } from 'https://cdn.jsdelivr.net/npm/@huggingface/[email protected]/+esm';
+    import { InferenceClient } from 'https://cdn.jsdelivr.net/npm/@huggingface/[email protected]/+esm';
     import { createRepo, commit, deleteRepo, listFiles } from "https://cdn.jsdelivr.net/npm/@huggingface/[email protected]/+esm";
 </script>
 ```
@@ -105,12 +106,12 @@ You can run our packages with vanilla JS, without any bundler, by using a CDN or
 
 ```ts
 // esm.sh
-import { HfInference } from "https://esm.sh/@huggingface/inference"
+import { InferenceClient } from "https://esm.sh/@huggingface/inference"
 import { HfAgent } from "https://esm.sh/@huggingface/agents";
 
 import { createRepo, commit, deleteRepo, listFiles } from "https://esm.sh/@huggingface/hub"
 // or npm:
-import { HfInference } from "npm:@huggingface/inference"
+import { InferenceClient } from "npm:@huggingface/inference"
 import { HfAgent } from "npm:@huggingface/agents";
 
 import { createRepo, commit, deleteRepo, listFiles } from "npm:@huggingface/hub"
@@ -123,11 +124,11 @@ Get your HF access token in your [account settings](https://huggingface.co/setti
 ### @huggingface/inference examples
 
 ```ts
-import { HfInference } from "@huggingface/inference";
+import { InferenceClient } from "@huggingface/inference";
 
 const HF_TOKEN = "hf_...";
 
-const inference = new HfInference(HF_TOKEN);
+const inference = new InferenceClient(HF_TOKEN);
 
 // Chat completion API
 const out = await inference.chatCompletion({
@@ -179,7 +180,7 @@ await inference.imageToText({
 
 // Using your own dedicated inference endpoint: https://hf.co/docs/inference-endpoints/
 const gpt2 = inference.endpoint('https://xyz.eu-west-1.aws.endpoints.huggingface.cloud/gpt2');
-const { generated_text } = await gpt2.textGeneration({inputs: 'The answer to the universe is'});
+const { generated_text } = await gpt2.textGeneration({ inputs: 'The answer to the universe is' });
 
 // Chat Completion
 const llamaEndpoint = inference.endpoint(

e2e/deno/index.ts

@@ -1,4 +1,4 @@
-import { HfInference } from "npm:@huggingface/inference@*";
+import { InferenceClient } from "npm:@huggingface/inference@*";
 import { whoAmI, listFiles } from "npm:@huggingface/hub@*";
 
 const info = await whoAmI({ credentials: { accessToken: "hf_hub.js" }, hubUrl: "https://hub-ci.huggingface.co" });
@@ -10,7 +10,7 @@ for await (const file of listFiles({ repo: "gpt2" })) {
 
 const token = Deno.env.get("HF_TOKEN");
 if (token) {
-	const hf = new HfInference(token);
+	const hf = new InferenceClient(token);
 
 	const tokenInfo = await whoAmI({ credentials: { accessToken: token } });
 	console.log(tokenInfo);

e2e/svelte/src/routes/+page.svelte

@@ -1,8 +1,8 @@
 <script>
 	import { whoAmI, listFiles } from "@huggingface/hub";
-	import { HfInference } from "@huggingface/inference";
+	import { InferenceClient } from "@huggingface/inference";
 
-	const hf = new HfInference();
+	const hf = new InferenceClient();
 
 	const test = async () => {
 		const info = await whoAmI({ credentials: { accessToken: "hf_hub.js" }, hubUrl: "https://hub-ci.huggingface.co" });

e2e/ts/src/index.ts

@@ -1,9 +1,9 @@
-import { HfInference } from "@huggingface/inference";
+import { InferenceClient } from "@huggingface/inference";
 import { whoAmI } from "@huggingface/hub";
 
 const hfToken = process.env.token;
 
-const hf = new HfInference(hfToken);
+const hf = new InferenceClient(hfToken);
 
 (async () => {
 	const info = await whoAmI({ credentials: { accessToken: "hf_hub.js" }, hubUrl: "https://hub-ci.huggingface.co" });

packages/agents/package.json

@@ -56,6 +56,6 @@
 		"@types/node": "^18.13.0"
 	},
 	"dependencies": {
-		"@huggingface/inference": "^2.6.1"
+		"@huggingface/inference": "workspace:^"
 	}
 }

packages/agents/pnpm-lock.yaml

@@ -1,4 +1,4 @@
-import { HfInference } from "@huggingface/inference";
+import { InferenceClient } from "@huggingface/inference";
 import type { Data, Tool } from "../types";
 
 // this function passes the tools & files to the context before calling eval
@@ -17,7 +17,7 @@ export async function evalBuilder(
 
 		// add tools to context
 		for (const tool of tools) {
-			const toolCall = (input: Promise<Data>) => tool.call?.(input, new HfInference(accessToken ?? ""));
+			const toolCall = (input: Promise<Data>) => tool.call?.(input, new InferenceClient(accessToken ?? ""));
 			// @ts-expect-error adding to the scope
 			globalThis[tool.name] = toolCall;
 		}

packages/agents/src/lib/evalBuilder.ts

@@ -1,8 +1,8 @@
 import type { LLM } from "../types";
-import { HfInference } from "@huggingface/inference";
+import { InferenceClient } from "@huggingface/inference";
 
 export function LLMFromHub(accessToken?: string, model?: string): LLM {
-	const inference = new HfInference(accessToken);
+	const inference = new InferenceClient(accessToken);
 
 	return async (prompt: string): Promise<string> => {
 		const formattedPrompt = "<|user|>" + prompt + "<|end|><|assistant|>";
@@ -20,7 +20,7 @@ export function LLMFromHub(accessToken?: string, model?: string): LLM {
 }
 
 export function LLMFromEndpoint(accessToken: string, endpoint: string): LLM {
-	const inference = new HfInference(accessToken).endpoint(endpoint);
+	const inference = new InferenceClient(accessToken).endpoint(endpoint);
 	return async (prompt: string): Promise<string> => {
 		const formattedPrompt = "<|user|>" + prompt + "<|end|><|assistant|>";
 

packages/agents/src/llms/LLMHF.ts

@@ -15,9 +15,12 @@ export const imageToTextTool: Tool = {
 		if (typeof data === "string") throw "Input must be a blob.";
 
 		return (
-			await inference.imageToText({
-				data,
-			})
-		).generated_text;
+			// eslint-disable-next-line @typescript-eslint/no-non-null-assertion
+			(
+				await inference.imageToText({
+					data,
+				})
+			).generated_text!
+		);
 	},
 };

packages/agents/src/tools/imageToText.ts

@@ -1,12 +1,12 @@
-import type { HfInference } from "@huggingface/inference";
+import type { InferenceClient } from "@huggingface/inference";
 
 export type Data = string | Blob | ArrayBuffer;
 
 export interface Tool {
 	name: string;
 	description: string;
 	examples: Array<Example>;
-	call?: (input: Promise<Data>, inference: HfInference) => Promise<Data>;
+	call?: (input: Promise<Data>, inference: InferenceClient) => Promise<Data>;
 }
 
 export interface Example {