@huggingface/inference 3.1.5 → 3.2.0

package/README.md

@@ -1,6 +1,6 @@
 # 🤗 Hugging Face Inference
 
-A Typescript powered wrapper for the Hugging Face Inference API (serverless), Inference Endpoints (dedicated), and third-party Inference Providers.
+A Typescript powered wrapper for the HF Inference API (serverless), Inference Endpoints (dedicated), and third-party Inference Providers.
 It works with [Inference API (serverless)](https://huggingface.co/docs/api-inference/index) and [Inference Endpoints (dedicated)](https://huggingface.co/docs/inference-endpoints/index), and even with supported third-party Inference Providers.
 
 Check out the [full documentation](https://huggingface.co/docs/huggingface.js/inference/README).
@@ -117,7 +117,7 @@ for await (const output of hf.textGenerationStream({
 
 ### Text Generation (Chat Completion API Compatible)
 
-Using the `chatCompletion` method, you can generate text with models compatible with the OpenAI Chat Completion API. All models served by [TGI](https://api-inference.huggingface.co/framework/text-generation-inference) on Hugging Face support Messages API.
+Using the `chatCompletion` method, you can generate text with models compatible with the OpenAI Chat Completion API. All models served by [TGI](https://huggingface.co/docs/text-generation-inference/) on Hugging Face support Messages API.
 
 [Demo](https://huggingface.co/spaces/huggingfacejs/streaming-chat-completion)
 
@@ -611,7 +611,7 @@ const { generated_text } = await gpt2.textGeneration({inputs: 'The answer to the
 
 // Chat Completion Example
 const ep = hf.endpoint(
-  "https://api-inference.huggingface.co/models/meta-llama/Llama-3.1-8B-Instruct"
+  "https://router.huggingface.co/hf-inference/models/meta-llama/Llama-3.1-8B-Instruct"
 );
 const stream = ep.chatCompletionStream({
   model: "tgi",

package/dist/index.cjs

@@ -20,14 +20,10 @@ var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: tru
 // src/index.ts
 var src_exports = {};
 __export(src_exports, {
-  FAL_AI_SUPPORTED_MODEL_IDS: () => FAL_AI_SUPPORTED_MODEL_IDS,
   HfInference: () => HfInference,
   HfInferenceEndpoint: () => HfInferenceEndpoint,
   INFERENCE_PROVIDERS: () => INFERENCE_PROVIDERS,
   InferenceOutputError: () => InferenceOutputError,
-  REPLICATE_SUPPORTED_MODEL_IDS: () => REPLICATE_SUPPORTED_MODEL_IDS,
-  SAMBANOVA_SUPPORTED_MODEL_IDS: () => SAMBANOVA_SUPPORTED_MODEL_IDS,
-  TOGETHER_SUPPORTED_MODEL_IDS: () => TOGETHER_SUPPORTED_MODEL_IDS,
   audioClassification: () => audioClassification,
   audioToAudio: () => audioToAudio,
   automaticSpeechRecognition: () => automaticSpeechRecognition,
@@ -102,126 +98,19 @@ __export(tasks_exports, {
 
 // src/config.ts
 var HF_HUB_URL = "https://huggingface.co";
+var HF_ROUTER_URL = "https://router.huggingface.co";
 
 // src/providers/fal-ai.ts
 var FAL_AI_API_BASE_URL = "https://fal.run";
-var FAL_AI_SUPPORTED_MODEL_IDS = {
-  "text-to-image": {
-    "black-forest-labs/FLUX.1-schnell": "fal-ai/flux/schnell",
-    "black-forest-labs/FLUX.1-dev": "fal-ai/flux/dev",
-    "playgroundai/playground-v2.5-1024px-aesthetic": "fal-ai/playground-v25",
-    "ByteDance/SDXL-Lightning": "fal-ai/lightning-models",
-    "PixArt-alpha/PixArt-Sigma-XL-2-1024-MS": "fal-ai/pixart-sigma",
-    "stabilityai/stable-diffusion-3-medium": "fal-ai/stable-diffusion-v3-medium",
-    "Warlord-K/Sana-1024": "fal-ai/sana",
-    "fal/AuraFlow-v0.2": "fal-ai/aura-flow",
-    "stabilityai/stable-diffusion-3.5-large": "fal-ai/stable-diffusion-v35-large",
-    "stabilityai/stable-diffusion-3.5-large-turbo": "fal-ai/stable-diffusion-v35-large/turbo",
-    "stabilityai/stable-diffusion-3.5-medium": "fal-ai/stable-diffusion-v35-medium",
-    "Kwai-Kolors/Kolors": "fal-ai/kolors"
-  },
-  "automatic-speech-recognition": {
-    "openai/whisper-large-v3": "fal-ai/whisper"
-  },
-  "text-to-video": {
-    "genmo/mochi-1-preview": "fal-ai/mochi-v1",
-    "tencent/HunyuanVideo": "fal-ai/hunyuan-video",
-    "THUDM/CogVideoX-5b": "fal-ai/cogvideox-5b",
-    "Lightricks/LTX-Video": "fal-ai/ltx-video"
-  }
-};
 
 // src/providers/replicate.ts
 var REPLICATE_API_BASE_URL = "https://api.replicate.com";
-var REPLICATE_SUPPORTED_MODEL_IDS = {
-  "text-to-image": {
-    "black-forest-labs/FLUX.1-dev": "black-forest-labs/flux-dev",
-    "black-forest-labs/FLUX.1-schnell": "black-forest-labs/flux-schnell",
-    "ByteDance/Hyper-SD": "bytedance/hyper-flux-16step:382cf8959fb0f0d665b26e7e80b8d6dc3faaef1510f14ce017e8c732bb3d1eb7",
-    "ByteDance/SDXL-Lightning": "bytedance/sdxl-lightning-4step:5599ed30703defd1d160a25a63321b4dec97101d98b4674bcc56e41f62f35637",
-    "playgroundai/playground-v2.5-1024px-aesthetic": "playgroundai/playground-v2.5-1024px-aesthetic:a45f82a1382bed5c7aeb861dac7c7d191b0fdf74d8d57c4a0e6ed7d4d0bf7d24",
-    "stabilityai/stable-diffusion-3.5-large-turbo": "stability-ai/stable-diffusion-3.5-large-turbo",
-    "stabilityai/stable-diffusion-3.5-large": "stability-ai/stable-diffusion-3.5-large",
-    "stabilityai/stable-diffusion-3.5-medium": "stability-ai/stable-diffusion-3.5-medium",
-    "stabilityai/stable-diffusion-xl-base-1.0": "stability-ai/sdxl:7762fd07cf82c948538e41f63f77d685e02b063e37e496e96eefd46c929f9bdc"
-  },
-  "text-to-speech": {
-    "OuteAI/OuteTTS-0.3-500M": "jbilcke/oute-tts:39a59319327b27327fa3095149c5a746e7f2aee18c75055c3368237a6503cd26"
-  },
-  "text-to-video": {
-    "genmo/mochi-1-preview": "genmoai/mochi-1:1944af04d098ef69bed7f9d335d102e652203f268ec4aaa2d836f6217217e460"
-  }
-};
 
 // src/providers/sambanova.ts
 var SAMBANOVA_API_BASE_URL = "https://api.sambanova.ai";
-var SAMBANOVA_SUPPORTED_MODEL_IDS = {
-  /** Chat completion / conversational */
-  conversational: {
-    "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
-    "Qwen/Qwen2.5-72B-Instruct": "Qwen2.5-72B-Instruct",
-    "Qwen/QwQ-32B-Preview": "QwQ-32B-Preview",
-    "meta-llama/Llama-3.3-70B-Instruct": "Meta-Llama-3.3-70B-Instruct",
-    "meta-llama/Llama-3.2-1B-Instruct": "Meta-Llama-3.2-1B-Instruct",
-    "meta-llama/Llama-3.2-3B-Instruct": "Meta-Llama-3.2-3B-Instruct",
-    "meta-llama/Llama-3.2-11B-Vision-Instruct": "Llama-3.2-11B-Vision-Instruct",
-    "meta-llama/Llama-3.2-90B-Vision-Instruct": "Llama-3.2-90B-Vision-Instruct",
-    "meta-llama/Llama-3.1-8B-Instruct": "Meta-Llama-3.1-8B-Instruct",
-    "meta-llama/Llama-3.1-70B-Instruct": "Meta-Llama-3.1-70B-Instruct",
-    "meta-llama/Llama-3.1-405B-Instruct": "Meta-Llama-3.1-405B-Instruct",
-    "meta-llama/Llama-Guard-3-8B": "Meta-Llama-Guard-3-8B"
-  }
-};
 
 // src/providers/together.ts
 var TOGETHER_API_BASE_URL = "https://api.together.xyz";
-var TOGETHER_SUPPORTED_MODEL_IDS = {
-  "text-to-image": {
-    "black-forest-labs/FLUX.1-Canny-dev": "black-forest-labs/FLUX.1-canny",
-    "black-forest-labs/FLUX.1-Depth-dev": "black-forest-labs/FLUX.1-depth",
-    "black-forest-labs/FLUX.1-dev": "black-forest-labs/FLUX.1-dev",
-    "black-forest-labs/FLUX.1-Redux-dev": "black-forest-labs/FLUX.1-redux",
-    "black-forest-labs/FLUX.1-schnell": "black-forest-labs/FLUX.1-pro",
-    "stabilityai/stable-diffusion-xl-base-1.0": "stabilityai/stable-diffusion-xl-base-1.0"
-  },
-  conversational: {
-    "databricks/dbrx-instruct": "databricks/dbrx-instruct",
-    "deepseek-ai/DeepSeek-R1": "deepseek-ai/DeepSeek-R1",
-    "deepseek-ai/DeepSeek-V3": "deepseek-ai/DeepSeek-V3",
-    "deepseek-ai/deepseek-llm-67b-chat": "deepseek-ai/deepseek-llm-67b-chat",
-    "google/gemma-2-9b-it": "google/gemma-2-9b-it",
-    "google/gemma-2b-it": "google/gemma-2-27b-it",
-    "meta-llama/Llama-2-13b-chat-hf": "meta-llama/Llama-2-13b-chat-hf",
-    "meta-llama/Llama-2-7b-chat-hf": "meta-llama/Llama-2-7b-chat-hf",
-    "meta-llama/Llama-3.2-11B-Vision-Instruct": "meta-llama/Llama-Vision-Free",
-    "meta-llama/Llama-3.2-3B-Instruct": "meta-llama/Llama-3.2-3B-Instruct-Turbo",
-    "meta-llama/Llama-3.2-90B-Vision-Instruct": "meta-llama/Llama-3.2-90B-Vision-Instruct-Turbo",
-    "meta-llama/Llama-3.3-70B-Instruct": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
-    "meta-llama/Meta-Llama-3-70B-Instruct": "meta-llama/Llama-3-70b-chat-hf",
-    "meta-llama/Meta-Llama-3-8B-Instruct": "meta-llama/Meta-Llama-3-8B-Instruct-Turbo",
-    "meta-llama/Meta-Llama-3.1-405B-Instruct": "meta-llama/Llama-3.2-11B-Vision-Instruct-Turbo",
-    "meta-llama/Meta-Llama-3.1-70B-Instruct": "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo",
-    "meta-llama/Meta-Llama-3.1-8B-Instruct": "meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo-128K",
-    "microsoft/WizardLM-2-8x22B": "microsoft/WizardLM-2-8x22B",
-    "mistralai/Mistral-7B-Instruct-v0.3": "mistralai/Mistral-7B-Instruct-v0.3",
-    "mistralai/Mistral-Small-24B-Instruct-2501": "mistralai/Mistral-Small-24B-Instruct-2501",
-    "mistralai/Mixtral-8x22B-Instruct-v0.1": "mistralai/Mixtral-8x22B-Instruct-v0.1",
-    "mistralai/Mixtral-8x7B-Instruct-v0.1": "mistralai/Mixtral-8x7B-Instruct-v0.1",
-    "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
-    "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF": "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF",
-    "Qwen/Qwen2-72B-Instruct": "Qwen/Qwen2-72B-Instruct",
-    "Qwen/Qwen2.5-72B-Instruct": "Qwen/Qwen2.5-72B-Instruct-Turbo",
-    "Qwen/Qwen2.5-7B-Instruct": "Qwen/Qwen2.5-7B-Instruct-Turbo",
-    "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen/Qwen2.5-Coder-32B-Instruct",
-    "Qwen/QwQ-32B-Preview": "Qwen/QwQ-32B-Preview",
-    "scb10x/llama-3-typhoon-v1.5-8b-instruct": "scb10x/scb10x-llama3-typhoon-v1-5-8b-instruct",
-    "scb10x/llama-3-typhoon-v1.5x-70b-instruct-awq": "scb10x/scb10x-llama3-typhoon-v1-5x-4f316"
-  },
-  "text-generation": {
-    "meta-llama/Llama-2-70b-hf": "meta-llama/Llama-2-70b-hf",
-    "mistralai/Mixtral-8x7B-v0.1": "mistralai/Mixtral-8x7B-v0.1"
-  }
-};
 
 // src/lib/isUrl.ts
 function isUrl(modelOrUrl) {
@@ -230,10 +119,62 @@ function isUrl(modelOrUrl) {
 
 // package.json
 var name = "@huggingface/inference";
-var version = "3.1.5";
+var version = "3.2.0";
+
+// src/providers/consts.ts
+var HARDCODED_MODEL_ID_MAPPING = {
+  /**
+   * "HF model ID" => "Model ID on Inference Provider's side"
+   */
+  // "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
+};
+
+// src/lib/getProviderModelId.ts
+var inferenceProviderMappingCache = /* @__PURE__ */ new Map();
+async function getProviderModelId(params, args, options = {}) {
+  if (params.provider === "hf-inference") {
+    return params.model;
+  }
+  if (!options.taskHint) {
+    throw new Error("taskHint must be specified when using a third-party provider");
+  }
+  const task = options.taskHint === "text-generation" && options.chatCompletion ? "conversational" : options.taskHint;
+  if (HARDCODED_MODEL_ID_MAPPING[params.model]) {
+    return HARDCODED_MODEL_ID_MAPPING[params.model];
+  }
+  let inferenceProviderMapping;
+  if (inferenceProviderMappingCache.has(params.model)) {
+    inferenceProviderMapping = inferenceProviderMappingCache.get(params.model);
+  } else {
+    inferenceProviderMapping = await (options?.fetch ?? fetch)(
+      `${HF_HUB_URL}/api/models/${params.model}?expand[]=inferenceProviderMapping`,
+      {
+        headers: args.accessToken?.startsWith("hf_") ? { Authorization: `Bearer ${args.accessToken}` } : {}
+      }
+    ).then((resp) => resp.json()).then((json) => json.inferenceProviderMapping).catch(() => null);
+  }
+  if (!inferenceProviderMapping) {
+    throw new Error(`We have not been able to find inference provider information for model ${params.model}.`);
+  }
+  const providerMapping = inferenceProviderMapping[params.provider];
+  if (providerMapping) {
+    if (providerMapping.task !== task) {
+      throw new Error(
+        `Model ${params.model} is not supported for task ${task} and provider ${params.provider}. Supported task: ${providerMapping.task}.`
+      );
+    }
+    if (providerMapping.status === "staging") {
+      console.warn(
+        `Model ${params.model} is in staging mode for provider ${params.provider}. Meant for test purposes only.`
+      );
+    }
+    return providerMapping.providerId;
+  }
+  throw new Error(`Model ${params.model} is not supported provider ${params.provider}.`);
+}
 
 // src/lib/makeRequestOptions.ts
-var HF_HUB_INFERENCE_PROXY_TEMPLATE = `${HF_HUB_URL}/api/inference-proxy/{{PROVIDER}}`;
+var HF_HUB_INFERENCE_PROXY_TEMPLATE = `${HF_ROUTER_URL}/{{PROVIDER}}`;
 var tasks = null;
 async function makeRequestOptions(args, options) {
   const { accessToken, endpointUrl, provider: maybeProvider, model: maybeModel, ...remainingArgs } = args;
@@ -249,16 +190,15 @@ async function makeRequestOptions(args, options) {
   if (maybeModel && isUrl(maybeModel)) {
     throw new Error(`Model URLs are no longer supported. Use endpointUrl instead.`);
   }
-  let model;
-  if (!maybeModel) {
-    if (taskHint) {
-      model = mapModel({ model: await loadDefaultModel(taskHint), provider, taskHint, chatCompletion: chatCompletion2 });
-    } else {
-      throw new Error("No model provided, and no default model found for this task");
-    }
-  } else {
-    model = mapModel({ model: maybeModel, provider, taskHint, chatCompletion: chatCompletion2 });
+  if (!maybeModel && !taskHint) {
+    throw new Error("No model provided, and no task has been specified.");
   }
+  const hfModel = maybeModel ?? await loadDefaultModel(taskHint);
+  const model = await getProviderModelId({ model: hfModel, provider }, args, {
+    taskHint,
+    chatCompletion: chatCompletion2,
+    fetch: options?.fetch
+  });
   const authMethod = accessToken ? accessToken.startsWith("hf_") ? "hf-token" : "provider-key" : includeCredentials === "include" ? "credentials-include" : "none";
   const url = endpointUrl ? chatCompletion2 ? endpointUrl + `/v1/chat/completions` : endpointUrl : makeUrl({
     authMethod,
@@ -314,31 +254,6 @@ async function makeRequestOptions(args, options) {
   };
   return { url, info };
 }
-function mapModel(params) {
-  if (params.provider === "hf-inference") {
-    return params.model;
-  }
-  if (!params.taskHint) {
-    throw new Error("taskHint must be specified when using a third-party provider");
-  }
-  const task = params.taskHint === "text-generation" && params.chatCompletion ? "conversational" : params.taskHint;
-  const model = (() => {
-    switch (params.provider) {
-      case "fal-ai":
-        return FAL_AI_SUPPORTED_MODEL_IDS[task]?.[params.model];
-      case "replicate":
-        return REPLICATE_SUPPORTED_MODEL_IDS[task]?.[params.model];
-      case "sambanova":
-        return SAMBANOVA_SUPPORTED_MODEL_IDS[task]?.[params.model];
-      case "together":
-        return TOGETHER_SUPPORTED_MODEL_IDS[task]?.[params.model];
-    }
-  })();
-  if (!model) {
-    throw new Error(`Model ${params.model} is not supported for task ${task} and provider ${params.provider}`);
-  }
-  return model;
-}
 function makeUrl(params) {
   if (params.authMethod === "none" && params.provider !== "hf-inference") {
     throw new Error("Authentication is required when requesting a third-party provider. Please provide accessToken");
@@ -722,7 +637,12 @@ async function buildPayload(args) {
 
 // src/tasks/audio/textToSpeech.ts
 async function textToSpeech(args, options) {
-  const res = await request(args, {
+  const payload = args.provider === "replicate" ? {
+    ...omit(args, ["inputs", "parameters"]),
+    ...args.parameters,
+    text: args.inputs
+  } : args;
+  const res = await request(payload, {
     ...options,
     taskHint: "text-to-speech"
   });
@@ -1352,14 +1272,10 @@ var HfInferenceEndpoint = class {
 var INFERENCE_PROVIDERS = ["fal-ai", "replicate", "sambanova", "together", "hf-inference"];
 // Annotate the CommonJS export names for ESM import in node:
 0 && (module.exports = {
-  FAL_AI_SUPPORTED_MODEL_IDS,
   HfInference,
   HfInferenceEndpoint,
   INFERENCE_PROVIDERS,
   InferenceOutputError,
-  REPLICATE_SUPPORTED_MODEL_IDS,
-  SAMBANOVA_SUPPORTED_MODEL_IDS,
-  TOGETHER_SUPPORTED_MODEL_IDS,
   audioClassification,
   audioToAudio,
   automaticSpeechRecognition,

package/dist/index.js

@@ -43,126 +43,19 @@ __export(tasks_exports, {
 
 // src/config.ts
 var HF_HUB_URL = "https://huggingface.co";
+var HF_ROUTER_URL = "https://router.huggingface.co";
 
 // src/providers/fal-ai.ts
 var FAL_AI_API_BASE_URL = "https://fal.run";
-var FAL_AI_SUPPORTED_MODEL_IDS = {
-  "text-to-image": {
-    "black-forest-labs/FLUX.1-schnell": "fal-ai/flux/schnell",
-    "black-forest-labs/FLUX.1-dev": "fal-ai/flux/dev",
-    "playgroundai/playground-v2.5-1024px-aesthetic": "fal-ai/playground-v25",
-    "ByteDance/SDXL-Lightning": "fal-ai/lightning-models",
-    "PixArt-alpha/PixArt-Sigma-XL-2-1024-MS": "fal-ai/pixart-sigma",
-    "stabilityai/stable-diffusion-3-medium": "fal-ai/stable-diffusion-v3-medium",
-    "Warlord-K/Sana-1024": "fal-ai/sana",
-    "fal/AuraFlow-v0.2": "fal-ai/aura-flow",
-    "stabilityai/stable-diffusion-3.5-large": "fal-ai/stable-diffusion-v35-large",
-    "stabilityai/stable-diffusion-3.5-large-turbo": "fal-ai/stable-diffusion-v35-large/turbo",
-    "stabilityai/stable-diffusion-3.5-medium": "fal-ai/stable-diffusion-v35-medium",
-    "Kwai-Kolors/Kolors": "fal-ai/kolors"
-  },
-  "automatic-speech-recognition": {
-    "openai/whisper-large-v3": "fal-ai/whisper"
-  },
-  "text-to-video": {
-    "genmo/mochi-1-preview": "fal-ai/mochi-v1",
-    "tencent/HunyuanVideo": "fal-ai/hunyuan-video",
-    "THUDM/CogVideoX-5b": "fal-ai/cogvideox-5b",
-    "Lightricks/LTX-Video": "fal-ai/ltx-video"
-  }
-};
 
 // src/providers/replicate.ts
 var REPLICATE_API_BASE_URL = "https://api.replicate.com";
-var REPLICATE_SUPPORTED_MODEL_IDS = {
-  "text-to-image": {
-    "black-forest-labs/FLUX.1-dev": "black-forest-labs/flux-dev",
-    "black-forest-labs/FLUX.1-schnell": "black-forest-labs/flux-schnell",
-    "ByteDance/Hyper-SD": "bytedance/hyper-flux-16step:382cf8959fb0f0d665b26e7e80b8d6dc3faaef1510f14ce017e8c732bb3d1eb7",
-    "ByteDance/SDXL-Lightning": "bytedance/sdxl-lightning-4step:5599ed30703defd1d160a25a63321b4dec97101d98b4674bcc56e41f62f35637",
-    "playgroundai/playground-v2.5-1024px-aesthetic": "playgroundai/playground-v2.5-1024px-aesthetic:a45f82a1382bed5c7aeb861dac7c7d191b0fdf74d8d57c4a0e6ed7d4d0bf7d24",
-    "stabilityai/stable-diffusion-3.5-large-turbo": "stability-ai/stable-diffusion-3.5-large-turbo",
-    "stabilityai/stable-diffusion-3.5-large": "stability-ai/stable-diffusion-3.5-large",
-    "stabilityai/stable-diffusion-3.5-medium": "stability-ai/stable-diffusion-3.5-medium",
-    "stabilityai/stable-diffusion-xl-base-1.0": "stability-ai/sdxl:7762fd07cf82c948538e41f63f77d685e02b063e37e496e96eefd46c929f9bdc"
-  },
-  "text-to-speech": {
-    "OuteAI/OuteTTS-0.3-500M": "jbilcke/oute-tts:39a59319327b27327fa3095149c5a746e7f2aee18c75055c3368237a6503cd26"
-  },
-  "text-to-video": {
-    "genmo/mochi-1-preview": "genmoai/mochi-1:1944af04d098ef69bed7f9d335d102e652203f268ec4aaa2d836f6217217e460"
-  }
-};
 
 // src/providers/sambanova.ts
 var SAMBANOVA_API_BASE_URL = "https://api.sambanova.ai";
-var SAMBANOVA_SUPPORTED_MODEL_IDS = {
-  /** Chat completion / conversational */
-  conversational: {
-    "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
-    "Qwen/Qwen2.5-72B-Instruct": "Qwen2.5-72B-Instruct",
-    "Qwen/QwQ-32B-Preview": "QwQ-32B-Preview",
-    "meta-llama/Llama-3.3-70B-Instruct": "Meta-Llama-3.3-70B-Instruct",
-    "meta-llama/Llama-3.2-1B-Instruct": "Meta-Llama-3.2-1B-Instruct",
-    "meta-llama/Llama-3.2-3B-Instruct": "Meta-Llama-3.2-3B-Instruct",
-    "meta-llama/Llama-3.2-11B-Vision-Instruct": "Llama-3.2-11B-Vision-Instruct",
-    "meta-llama/Llama-3.2-90B-Vision-Instruct": "Llama-3.2-90B-Vision-Instruct",
-    "meta-llama/Llama-3.1-8B-Instruct": "Meta-Llama-3.1-8B-Instruct",
-    "meta-llama/Llama-3.1-70B-Instruct": "Meta-Llama-3.1-70B-Instruct",
-    "meta-llama/Llama-3.1-405B-Instruct": "Meta-Llama-3.1-405B-Instruct",
-    "meta-llama/Llama-Guard-3-8B": "Meta-Llama-Guard-3-8B"
-  }
-};
 
 // src/providers/together.ts
 var TOGETHER_API_BASE_URL = "https://api.together.xyz";
-var TOGETHER_SUPPORTED_MODEL_IDS = {
-  "text-to-image": {
-    "black-forest-labs/FLUX.1-Canny-dev": "black-forest-labs/FLUX.1-canny",
-    "black-forest-labs/FLUX.1-Depth-dev": "black-forest-labs/FLUX.1-depth",
-    "black-forest-labs/FLUX.1-dev": "black-forest-labs/FLUX.1-dev",
-    "black-forest-labs/FLUX.1-Redux-dev": "black-forest-labs/FLUX.1-redux",
-    "black-forest-labs/FLUX.1-schnell": "black-forest-labs/FLUX.1-pro",
-    "stabilityai/stable-diffusion-xl-base-1.0": "stabilityai/stable-diffusion-xl-base-1.0"
-  },
-  conversational: {
-    "databricks/dbrx-instruct": "databricks/dbrx-instruct",
-    "deepseek-ai/DeepSeek-R1": "deepseek-ai/DeepSeek-R1",
-    "deepseek-ai/DeepSeek-V3": "deepseek-ai/DeepSeek-V3",
-    "deepseek-ai/deepseek-llm-67b-chat": "deepseek-ai/deepseek-llm-67b-chat",
-    "google/gemma-2-9b-it": "google/gemma-2-9b-it",
-    "google/gemma-2b-it": "google/gemma-2-27b-it",
-    "meta-llama/Llama-2-13b-chat-hf": "meta-llama/Llama-2-13b-chat-hf",
-    "meta-llama/Llama-2-7b-chat-hf": "meta-llama/Llama-2-7b-chat-hf",
-    "meta-llama/Llama-3.2-11B-Vision-Instruct": "meta-llama/Llama-Vision-Free",
-    "meta-llama/Llama-3.2-3B-Instruct": "meta-llama/Llama-3.2-3B-Instruct-Turbo",
-    "meta-llama/Llama-3.2-90B-Vision-Instruct": "meta-llama/Llama-3.2-90B-Vision-Instruct-Turbo",
-    "meta-llama/Llama-3.3-70B-Instruct": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
-    "meta-llama/Meta-Llama-3-70B-Instruct": "meta-llama/Llama-3-70b-chat-hf",
-    "meta-llama/Meta-Llama-3-8B-Instruct": "meta-llama/Meta-Llama-3-8B-Instruct-Turbo",
-    "meta-llama/Meta-Llama-3.1-405B-Instruct": "meta-llama/Llama-3.2-11B-Vision-Instruct-Turbo",
-    "meta-llama/Meta-Llama-3.1-70B-Instruct": "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo",
-    "meta-llama/Meta-Llama-3.1-8B-Instruct": "meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo-128K",
-    "microsoft/WizardLM-2-8x22B": "microsoft/WizardLM-2-8x22B",
-    "mistralai/Mistral-7B-Instruct-v0.3": "mistralai/Mistral-7B-Instruct-v0.3",
-    "mistralai/Mistral-Small-24B-Instruct-2501": "mistralai/Mistral-Small-24B-Instruct-2501",
-    "mistralai/Mixtral-8x22B-Instruct-v0.1": "mistralai/Mixtral-8x22B-Instruct-v0.1",
-    "mistralai/Mixtral-8x7B-Instruct-v0.1": "mistralai/Mixtral-8x7B-Instruct-v0.1",
-    "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
-    "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF": "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF",
-    "Qwen/Qwen2-72B-Instruct": "Qwen/Qwen2-72B-Instruct",
-    "Qwen/Qwen2.5-72B-Instruct": "Qwen/Qwen2.5-72B-Instruct-Turbo",
-    "Qwen/Qwen2.5-7B-Instruct": "Qwen/Qwen2.5-7B-Instruct-Turbo",
-    "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen/Qwen2.5-Coder-32B-Instruct",
-    "Qwen/QwQ-32B-Preview": "Qwen/QwQ-32B-Preview",
-    "scb10x/llama-3-typhoon-v1.5-8b-instruct": "scb10x/scb10x-llama3-typhoon-v1-5-8b-instruct",
-    "scb10x/llama-3-typhoon-v1.5x-70b-instruct-awq": "scb10x/scb10x-llama3-typhoon-v1-5x-4f316"
-  },
-  "text-generation": {
-    "meta-llama/Llama-2-70b-hf": "meta-llama/Llama-2-70b-hf",
-    "mistralai/Mixtral-8x7B-v0.1": "mistralai/Mixtral-8x7B-v0.1"
-  }
-};
 
 // src/lib/isUrl.ts
 function isUrl(modelOrUrl) {
@@ -171,10 +64,62 @@ function isUrl(modelOrUrl) {
 
 // package.json
 var name = "@huggingface/inference";
-var version = "3.1.5";
+var version = "3.2.0";
+
+// src/providers/consts.ts
+var HARDCODED_MODEL_ID_MAPPING = {
+  /**
+   * "HF model ID" => "Model ID on Inference Provider's side"
+   */
+  // "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
+};
+
+// src/lib/getProviderModelId.ts
+var inferenceProviderMappingCache = /* @__PURE__ */ new Map();
+async function getProviderModelId(params, args, options = {}) {
+  if (params.provider === "hf-inference") {
+    return params.model;
+  }
+  if (!options.taskHint) {
+    throw new Error("taskHint must be specified when using a third-party provider");
+  }
+  const task = options.taskHint === "text-generation" && options.chatCompletion ? "conversational" : options.taskHint;
+  if (HARDCODED_MODEL_ID_MAPPING[params.model]) {
+    return HARDCODED_MODEL_ID_MAPPING[params.model];
+  }
+  let inferenceProviderMapping;
+  if (inferenceProviderMappingCache.has(params.model)) {
+    inferenceProviderMapping = inferenceProviderMappingCache.get(params.model);
+  } else {
+    inferenceProviderMapping = await (options?.fetch ?? fetch)(
+      `${HF_HUB_URL}/api/models/${params.model}?expand[]=inferenceProviderMapping`,
+      {
+        headers: args.accessToken?.startsWith("hf_") ? { Authorization: `Bearer ${args.accessToken}` } : {}
+      }
+    ).then((resp) => resp.json()).then((json) => json.inferenceProviderMapping).catch(() => null);
+  }
+  if (!inferenceProviderMapping) {
+    throw new Error(`We have not been able to find inference provider information for model ${params.model}.`);
+  }
+  const providerMapping = inferenceProviderMapping[params.provider];
+  if (providerMapping) {
+    if (providerMapping.task !== task) {
+      throw new Error(
+        `Model ${params.model} is not supported for task ${task} and provider ${params.provider}. Supported task: ${providerMapping.task}.`
+      );
+    }
+    if (providerMapping.status === "staging") {
+      console.warn(
+        `Model ${params.model} is in staging mode for provider ${params.provider}. Meant for test purposes only.`
+      );
+    }
+    return providerMapping.providerId;
+  }
+  throw new Error(`Model ${params.model} is not supported provider ${params.provider}.`);
+}
 
 // src/lib/makeRequestOptions.ts
-var HF_HUB_INFERENCE_PROXY_TEMPLATE = `${HF_HUB_URL}/api/inference-proxy/{{PROVIDER}}`;
+var HF_HUB_INFERENCE_PROXY_TEMPLATE = `${HF_ROUTER_URL}/{{PROVIDER}}`;
 var tasks = null;
 async function makeRequestOptions(args, options) {
   const { accessToken, endpointUrl, provider: maybeProvider, model: maybeModel, ...remainingArgs } = args;
@@ -190,16 +135,15 @@ async function makeRequestOptions(args, options) {
   if (maybeModel && isUrl(maybeModel)) {
     throw new Error(`Model URLs are no longer supported. Use endpointUrl instead.`);
   }
-  let model;
-  if (!maybeModel) {
-    if (taskHint) {
-      model = mapModel({ model: await loadDefaultModel(taskHint), provider, taskHint, chatCompletion: chatCompletion2 });
-    } else {
-      throw new Error("No model provided, and no default model found for this task");
-    }
-  } else {
-    model = mapModel({ model: maybeModel, provider, taskHint, chatCompletion: chatCompletion2 });
+  if (!maybeModel && !taskHint) {
+    throw new Error("No model provided, and no task has been specified.");
   }
+  const hfModel = maybeModel ?? await loadDefaultModel(taskHint);
+  const model = await getProviderModelId({ model: hfModel, provider }, args, {
+    taskHint,
+    chatCompletion: chatCompletion2,
+    fetch: options?.fetch
+  });
   const authMethod = accessToken ? accessToken.startsWith("hf_") ? "hf-token" : "provider-key" : includeCredentials === "include" ? "credentials-include" : "none";
   const url = endpointUrl ? chatCompletion2 ? endpointUrl + `/v1/chat/completions` : endpointUrl : makeUrl({
     authMethod,
@@ -255,31 +199,6 @@ async function makeRequestOptions(args, options) {
   };
   return { url, info };
 }
-function mapModel(params) {
-  if (params.provider === "hf-inference") {
-    return params.model;
-  }
-  if (!params.taskHint) {
-    throw new Error("taskHint must be specified when using a third-party provider");
-  }
-  const task = params.taskHint === "text-generation" && params.chatCompletion ? "conversational" : params.taskHint;
-  const model = (() => {
-    switch (params.provider) {
-      case "fal-ai":
-        return FAL_AI_SUPPORTED_MODEL_IDS[task]?.[params.model];
-      case "replicate":
-        return REPLICATE_SUPPORTED_MODEL_IDS[task]?.[params.model];
-      case "sambanova":
-        return SAMBANOVA_SUPPORTED_MODEL_IDS[task]?.[params.model];
-      case "together":
-        return TOGETHER_SUPPORTED_MODEL_IDS[task]?.[params.model];
-    }
-  })();
-  if (!model) {
-    throw new Error(`Model ${params.model} is not supported for task ${task} and provider ${params.provider}`);
-  }
-  return model;
-}
 function makeUrl(params) {
   if (params.authMethod === "none" && params.provider !== "hf-inference") {
     throw new Error("Authentication is required when requesting a third-party provider. Please provide accessToken");
@@ -663,7 +582,12 @@ async function buildPayload(args) {
 
 // src/tasks/audio/textToSpeech.ts
 async function textToSpeech(args, options) {
-  const res = await request(args, {
+  const payload = args.provider === "replicate" ? {
+    ...omit(args, ["inputs", "parameters"]),
+    ...args.parameters,
+    text: args.inputs
+  } : args;
+  const res = await request(payload, {
     ...options,
     taskHint: "text-to-speech"
   });
@@ -1292,14 +1216,10 @@ var HfInferenceEndpoint = class {
 // src/types.ts
 var INFERENCE_PROVIDERS = ["fal-ai", "replicate", "sambanova", "together", "hf-inference"];
 export {
-  FAL_AI_SUPPORTED_MODEL_IDS,
   HfInference,
   HfInferenceEndpoint,
   INFERENCE_PROVIDERS,
   InferenceOutputError,
-  REPLICATE_SUPPORTED_MODEL_IDS,
-  SAMBANOVA_SUPPORTED_MODEL_IDS,
-  TOGETHER_SUPPORTED_MODEL_IDS,
   audioClassification,
   audioToAudio,
   automaticSpeechRecognition,

package/dist/src/config.d.ts

@@ -1,2 +1,3 @@
 export declare const HF_HUB_URL = "https://huggingface.co";
+export declare const HF_ROUTER_URL = "https://router.huggingface.co";
 //# sourceMappingURL=config.d.ts.map

package/dist/src/config.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"config.d.ts","sourceRoot":"","sources":["../../src/config.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,UAAU,2BAA2B,CAAC"}
+{"version":3,"file":"config.d.ts","sourceRoot":"","sources":["../../src/config.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,UAAU,2BAA2B,CAAC;AACnD,eAAO,MAAM,aAAa,kCAAkC,CAAC"}

package/dist/src/index.d.ts

@@ -1,10 +1,5 @@
-export type { ProviderMapping } from "./providers/types";
 export { HfInference, HfInferenceEndpoint } from "./HfInference";
 export { InferenceOutputError } from "./lib/InferenceOutputError";
-export { FAL_AI_SUPPORTED_MODEL_IDS } from "./providers/fal-ai";
-export { REPLICATE_SUPPORTED_MODEL_IDS } from "./providers/replicate";
-export { SAMBANOVA_SUPPORTED_MODEL_IDS } from "./providers/sambanova";
-export { TOGETHER_SUPPORTED_MODEL_IDS } from "./providers/together";
 export * from "./types";
 export * from "./tasks";
 //# sourceMappingURL=index.d.ts.map

package/dist/src/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA,YAAY,EAAE,eAAe,EAAE,MAAM,mBAAmB,CAAC;AACzD,OAAO,EAAE,WAAW,EAAE,mBAAmB,EAAE,MAAM,eAAe,CAAC;AACjE,OAAO,EAAE,oBAAoB,EAAE,MAAM,4BAA4B,CAAC;AAClE,OAAO,EAAE,0BAA0B,EAAE,MAAM,oBAAoB,CAAC;AAChE,OAAO,EAAE,6BAA6B,EAAE,MAAM,uBAAuB,CAAC;AACtE,OAAO,EAAE,6BAA6B,EAAE,MAAM,uBAAuB,CAAC;AACtE,OAAO,EAAE,4BAA4B,EAAE,MAAM,sBAAsB,CAAC;AACpE,cAAc,SAAS,CAAC;AACxB,cAAc,SAAS,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,mBAAmB,EAAE,MAAM,eAAe,CAAC;AACjE,OAAO,EAAE,oBAAoB,EAAE,MAAM,4BAA4B,CAAC;AAClE,cAAc,SAAS,CAAC;AACxB,cAAc,SAAS,CAAC"}

package/dist/src/lib/getProviderModelId.d.ts

@@ -0,0 +1,10 @@
+import type { InferenceProvider, InferenceTask, Options, RequestArgs } from "../types";
+export declare function getProviderModelId(params: {
+    model: string;
+    provider: InferenceProvider;
+}, args: RequestArgs, options?: {
+    taskHint?: InferenceTask;
+    chatCompletion?: boolean;
+    fetch?: Options["fetch"];
+}): Promise<string>;
+//# sourceMappingURL=getProviderModelId.d.ts.map

package/dist/src/lib/getProviderModelId.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"getProviderModelId.d.ts","sourceRoot":"","sources":["../../../src/lib/getProviderModelId.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,iBAAiB,EAAE,aAAa,EAAW,OAAO,EAAE,WAAW,EAAE,MAAM,UAAU,CAAC;AAShG,wBAAsB,kBAAkB,CACvC,MAAM,EAAE;IACP,KAAK,EAAE,MAAM,CAAC;IACd,QAAQ,EAAE,iBAAiB,CAAC;CAC5B,EACD,IAAI,EAAE,WAAW,EACjB,OAAO,GAAE;IACR,QAAQ,CAAC,EAAE,aAAa,CAAC;IACzB,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB,KAAK,CAAC,EAAE,OAAO,CAAC,OAAO,CAAC,CAAC;CACpB,GACJ,OAAO,CAAC,MAAM,CAAC,CAoDjB"}

package/dist/src/lib/makeRequestOptions.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"makeRequestOptions.d.ts","sourceRoot":"","sources":["../../../src/lib/makeRequestOptions.ts"],"names":[],"mappings":"AAOA,OAAO,KAAK,EAAE,aAAa,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,UAAU,CAAC;AAYpE;;GAEG;AACH,wBAAsB,kBAAkB,CACvC,IAAI,EAAE,WAAW,GAAG;IACnB,IAAI,CAAC,EAAE,IAAI,GAAG,WAAW,CAAC;IAC1B,MAAM,CAAC,EAAE,OAAO,CAAC;CACjB,EACD,OAAO,CAAC,EAAE,OAAO,GAAG;IACnB,yFAAyF;IACzF,SAAS,CAAC,EAAE,MAAM,GAAG,aAAa,CAAC;IACnC,sCAAsC;IACtC,QAAQ,CAAC,EAAE,aAAa,CAAC;IACzB,cAAc,CAAC,EAAE,OAAO,CAAC;CACzB,GACC,OAAO,CAAC;IAAE,GAAG,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,WAAW,CAAA;CAAE,CAAC,CAuH7C"}
+{"version":3,"file":"makeRequestOptions.d.ts","sourceRoot":"","sources":["../../../src/lib/makeRequestOptions.ts"],"names":[],"mappings":"AAMA,OAAO,KAAK,EAAE,aAAa,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,UAAU,CAAC;AAapE;;GAEG;AACH,wBAAsB,kBAAkB,CACvC,IAAI,EAAE,WAAW,GAAG;IACnB,IAAI,CAAC,EAAE,IAAI,GAAG,WAAW,CAAC;IAC1B,MAAM,CAAC,EAAE,OAAO,CAAC;CACjB,EACD,OAAO,CAAC,EAAE,OAAO,GAAG;IACnB,yFAAyF;IACzF,SAAS,CAAC,EAAE,MAAM,GAAG,aAAa,CAAC;IACnC,sCAAsC;IACtC,QAAQ,CAAC,EAAE,aAAa,CAAC;IACzB,cAAc,CAAC,EAAE,OAAO,CAAC;CACzB,GACC,OAAO,CAAC;IAAE,GAAG,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,WAAW,CAAA;CAAE,CAAC,CAqH7C"}

package/dist/src/providers/consts.d.ts

@@ -0,0 +1,10 @@
+import type { ModelId } from "../types";
+type ProviderId = string;
+/**
+ * If you want to try to run inference for a new model locally before it's registered on huggingface.co
+ * for a given Inference Provider,
+ * you can add it to the following dictionary, for dev purposes.
+ */
+export declare const HARDCODED_MODEL_ID_MAPPING: Record<ModelId, ProviderId>;
+export {};
+//# sourceMappingURL=consts.d.ts.map