@huggingface/inference - npm Package Compare versions

Comparing version 3.1.6 to 3.2.0

dist/index.js

@@ -46,124 +46,15 @@ var __defProp = Object.defineProperty;
 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:3c645149db020c85d080e2f8cfe482a0e68189a922cde964fa9e80fb179191f3",
-    "hexgrad/Kokoro-82M": "jaaari/kokoro-82m:dfdf537ba482b029e0a761699e6f55e9162cfd159270bfe0e44857caa5f275a6"
-  },
-  "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: {
-    "deepseek-ai/DeepSeek-Distill-R1-Llama-70B": "DeepSeek-Distill-R1-Llama-70B",
-    "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"
-  }
-};
 
@@ -177,6 +68,58 @@ // src/lib/isUrl.ts
 var name = "@huggingface/inference";
-var version = "3.1.6";
+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;
@@ -197,12 +140,11 @@ async function makeRequestOptions(args, options) {
   }
-  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";
@@ -263,27 +205,2 @@ const url = endpointUrl ? chatCompletion2 ? endpointUrl + `/v1/chat/completions` : endpointUrl : makeUrl({
 }
-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) {
@@ -1306,3 +1223,2 @@ if (params.authMethod === "none" && params.provider !== "hf-inference") {
 export {
-  FAL_AI_SUPPORTED_MODEL_IDS,
   HfInference,
@@ -1312,5 +1228,2 @@ HfInferenceEndpoint,
   InferenceOutputError,
-  REPLICATE_SUPPORTED_MODEL_IDS,
-  SAMBANOVA_SUPPORTED_MODEL_IDS,
-  TOGETHER_SUPPORTED_MODEL_IDS,
   audioClassification,
@@ -1317,0 +1230,0 @@ audioToAudio,

dist/src/config.d.ts

 export declare const HF_HUB_URL = "https://huggingface.co";
+export declare const HF_ROUTER_URL = "https://router.huggingface.co";
 //# sourceMappingURL=config.d.ts.map

dist/src/index.d.ts

@@ -1,10 +0,5 @@
-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

dist/src/providers/fal-ai.d.ts

@@ -1,6 +0,18 @@
-import type { ProviderMapping } from "./types";
 export declare const FAL_AI_API_BASE_URL = "https://fal.run";
-type FalAiId = string;
-export declare const FAL_AI_SUPPORTED_MODEL_IDS: ProviderMapping<FalAiId>;
-export {};
+/**
+ * See the registered mapping of HF model ID => Fal model ID here:
+ *
+ * https://huggingface.co/api/partners/fal-ai/models
+ *
+ * This is a publicly available mapping.
+ *
+ * If you want to try to run inference for a new model locally before it's registered on huggingface.co,
+ * you can add it to the dictionary "HARDCODED_MODEL_ID_MAPPING" in consts.ts, for dev purposes.
+ *
+ * - If you work at Fal and want to update this mapping, please use the model mapping API we provide on huggingface.co
+ * - If you're a community member and want to add a new supported HF model to Fal, please open an issue on the present repo
+ * and we will tag Fal team members.
+ *
+ * Thanks!
+ */
 //# sourceMappingURL=fal-ai.d.ts.map

dist/src/providers/replicate.d.ts

@@ -1,6 +0,18 @@
-import type { ProviderMapping } from "./types";
 export declare const REPLICATE_API_BASE_URL = "https://api.replicate.com";
-type ReplicateId = string;
-export declare const REPLICATE_SUPPORTED_MODEL_IDS: ProviderMapping<ReplicateId>;
-export {};
+/**
+ * See the registered mapping of HF model ID => Replicate model ID here:
+ *
+ * https://huggingface.co/api/partners/replicate/models
+ *
+ * This is a publicly available mapping.
+ *
+ * If you want to try to run inference for a new model locally before it's registered on huggingface.co,
+ * you can add it to the dictionary "HARDCODED_MODEL_ID_MAPPING" in consts.ts, for dev purposes.
+ *
+ * - If you work at Replicate and want to update this mapping, please use the model mapping API we provide on huggingface.co
+ * - If you're a community member and want to add a new supported HF model to Replicate, please open an issue on the present repo
+ * and we will tag Replicate team members.
+ *
+ * Thanks!
+ */
 //# sourceMappingURL=replicate.d.ts.map

dist/src/providers/sambanova.d.ts

@@ -1,6 +0,18 @@
-import type { ProviderMapping } from "./types";
 export declare const SAMBANOVA_API_BASE_URL = "https://api.sambanova.ai";
-type SambanovaId = string;
-export declare const SAMBANOVA_SUPPORTED_MODEL_IDS: ProviderMapping<SambanovaId>;
-export {};
+/**
+ * See the registered mapping of HF model ID => Sambanova model ID here:
+ *
+ * https://huggingface.co/api/partners/sambanova/models
+ *
+ * This is a publicly available mapping.
+ *
+ * If you want to try to run inference for a new model locally before it's registered on huggingface.co,
+ * you can add it to the dictionary "HARDCODED_MODEL_ID_MAPPING" in consts.ts, for dev purposes.
+ *
+ * - If you work at Sambanova and want to update this mapping, please use the model mapping API we provide on huggingface.co
+ * - If you're a community member and want to add a new supported HF model to Sambanova, please open an issue on the present repo
+ * and we will tag Sambanova team members.
+ *
+ * Thanks!
+ */
 //# sourceMappingURL=sambanova.d.ts.map

dist/src/providers/together.d.ts

@@ -1,12 +0,18 @@
-import type { ProviderMapping } from "./types";
 export declare const TOGETHER_API_BASE_URL = "https://api.together.xyz";
 /**
- * Same comment as in sambanova.ts
+ * See the registered mapping of HF model ID => Together model ID here:
+ *
+ * https://huggingface.co/api/partners/together/models
+ *
+ * This is a publicly available mapping.
+ *
+ * If you want to try to run inference for a new model locally before it's registered on huggingface.co,
+ * you can add it to the dictionary "HARDCODED_MODEL_ID_MAPPING" in consts.ts, for dev purposes.
+ *
+ * - If you work at Together and want to update this mapping, please use the model mapping API we provide on huggingface.co
+ * - If you're a community member and want to add a new supported HF model to Together, please open an issue on the present repo
+ * and we will tag Together team members.
+ *
+ * Thanks!
  */
-type TogetherId = string;
-/**
- * https://docs.together.ai/reference/models-1
- */
-export declare const TOGETHER_SUPPORTED_MODEL_IDS: ProviderMapping<TogetherId>;
-export {};
 //# sourceMappingURL=together.d.ts.map

package.json

 {
   "name": "@huggingface/inference",
-  "version": "3.1.6",
+  "version": "3.2.0",
   "packageManager": "pnpm@8.10.5",
@@ -42,3 +42,3 @@ "license": "MIT",
   "dependencies": {
-    "@huggingface/tasks": "^0.15.7"
+    "@huggingface/tasks": "^0.15.9"
   },
@@ -45,0 +45,0 @@ "devDependencies": {

README.md

@@ -120,3 +120,3 @@ # 🤗 Hugging Face Inference
 
-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.
 
@@ -615,3 +615,3 @@ [Demo](https://huggingface.co/spaces/huggingfacejs/streaming-chat-completion)
 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"
 );
@@ -618,0 +618,0 @@ const stream = ep.chatCompletionStream({

src/config.ts

 export const HF_HUB_URL = "https://huggingface.co";
+export const HF_ROUTER_URL = "https://router.huggingface.co";

src/index.ts

@@ -1,9 +0,4 @@
-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";

src/lib/makeRequestOptions.ts

@@ -1,7 +0,6 @@
-import type { WidgetType } from "@huggingface/tasks";
-import { HF_HUB_URL } from "../config";
-import { FAL_AI_API_BASE_URL, FAL_AI_SUPPORTED_MODEL_IDS } from "../providers/fal-ai";
-import { REPLICATE_API_BASE_URL, REPLICATE_SUPPORTED_MODEL_IDS } from "../providers/replicate";
-import { SAMBANOVA_API_BASE_URL, SAMBANOVA_SUPPORTED_MODEL_IDS } from "../providers/sambanova";
-import { TOGETHER_API_BASE_URL, TOGETHER_SUPPORTED_MODEL_IDS } from "../providers/together";
+import { HF_HUB_URL, HF_ROUTER_URL } from "../config";
+import { FAL_AI_API_BASE_URL } from "../providers/fal-ai";
+import { REPLICATE_API_BASE_URL } from "../providers/replicate";
+import { SAMBANOVA_API_BASE_URL } from "../providers/sambanova";
+import { TOGETHER_API_BASE_URL } from "../providers/together";
 import type { InferenceProvider } from "../types";
@@ -11,4 +10,5 @@ import type { InferenceTask, Options, RequestArgs } from "../types";
 import { version as packageVersion, name as packageName } from "../../package.json";
+import { getProviderModelId } from "./getProviderModelId";
 
-const HF_HUB_INFERENCE_PROXY_TEMPLATE = `${HF_HUB_URL}/api/inference-proxy/{{PROVIDER}}`;
+const HF_HUB_INFERENCE_PROXY_TEMPLATE = `${HF_ROUTER_URL}/{{PROVIDER}}`;
 
@@ -53,14 +53,12 @@ /**
 	}
-
-	let model: string;
-	if (!maybeModel) {
-		if (taskHint) {
-			model = mapModel({ model: await loadDefaultModel(taskHint), provider, taskHint, chatCompletion });
-		} else {
-			throw new Error("No model provided, and no default model found for this task");
-			/// TODO : change error message ^
-		}
-	} else {
-		model = mapModel({ model: maybeModel, provider, taskHint, chatCompletion });
+	if (!maybeModel && !taskHint) {
+		throw new Error("No model provided, and no task has been specified.");
 	}
+	// eslint-disable-next-line @typescript-eslint/no-non-null-assertion
+	const hfModel = maybeModel ?? (await loadDefaultModel(taskHint!));
+	const model = await getProviderModelId({ model: hfModel, provider }, args, {
+		taskHint,
+		chatCompletion,
+		fetch: options?.fetch,
+	});
 
@@ -158,35 +156,2 @@ /// If accessToken is passed, it should take precedence over includeCredentials
 
-function mapModel(params: {
-	model: string;
-	provider: InferenceProvider;
-	taskHint: InferenceTask | undefined;
-	chatCompletion: boolean | undefined;
-}): string {
-	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: WidgetType =
-		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: {
@@ -193,0 +158,0 @@ authMethod: "none" | "hf-token" | "credentials-include" | "provider-key";

src/providers/fal-ai.ts

@@ -1,31 +0,18 @@
-import type { ProviderMapping } from "./types";
-
 export const FAL_AI_API_BASE_URL = "https://fal.run";
 
-type FalAiId = string;
-
-export const FAL_AI_SUPPORTED_MODEL_IDS: ProviderMapping<FalAiId> = {
-	"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",
-	},
-};
+/**
+ * See the registered mapping of HF model ID => Fal model ID here:
+ *
+ * https://huggingface.co/api/partners/fal-ai/models
+ *
+ * This is a publicly available mapping.
+ *
+ * If you want to try to run inference for a new model locally before it's registered on huggingface.co,
+ * you can add it to the dictionary "HARDCODED_MODEL_ID_MAPPING" in consts.ts, for dev purposes.
+ *
+ * - If you work at Fal and want to update this mapping, please use the model mapping API we provide on huggingface.co
+ * - If you're a community member and want to add a new supported HF model to Fal, please open an issue on the present repo
+ * and we will tag Fal team members.
+ *
+ * Thanks!
+ */

src/providers/replicate.ts

@@ -1,30 +0,18 @@
-import type { ProviderMapping } from "./types";
-
 export const REPLICATE_API_BASE_URL = "https://api.replicate.com";
 
-type ReplicateId = string;
-
-export const REPLICATE_SUPPORTED_MODEL_IDS: ProviderMapping<ReplicateId> = {
-	"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:3c645149db020c85d080e2f8cfe482a0e68189a922cde964fa9e80fb179191f3",
-		"hexgrad/Kokoro-82M": "jaaari/kokoro-82m:dfdf537ba482b029e0a761699e6f55e9162cfd159270bfe0e44857caa5f275a6",
-	},
-	"text-to-video": {
-		"genmo/mochi-1-preview": "genmoai/mochi-1:1944af04d098ef69bed7f9d335d102e652203f268ec4aaa2d836f6217217e460",
-	},
-};
+/**
+ * See the registered mapping of HF model ID => Replicate model ID here:
+ *
+ * https://huggingface.co/api/partners/replicate/models
+ *
+ * This is a publicly available mapping.
+ *
+ * If you want to try to run inference for a new model locally before it's registered on huggingface.co,
+ * you can add it to the dictionary "HARDCODED_MODEL_ID_MAPPING" in consts.ts, for dev purposes.
+ *
+ * - If you work at Replicate and want to update this mapping, please use the model mapping API we provide on huggingface.co
+ * - If you're a community member and want to add a new supported HF model to Replicate, please open an issue on the present repo
+ * and we will tag Replicate team members.
+ *
+ * Thanks!
+ */

src/providers/sambanova.ts

@@ -1,24 +0,18 @@
-import type { ProviderMapping } from "./types";
-
 export const SAMBANOVA_API_BASE_URL = "https://api.sambanova.ai";
 
-type SambanovaId = string;
-
-export const SAMBANOVA_SUPPORTED_MODEL_IDS: ProviderMapping<SambanovaId> = {
-	/** Chat completion / conversational */
-	conversational: {
-		"deepseek-ai/DeepSeek-Distill-R1-Llama-70B": "DeepSeek-Distill-R1-Llama-70B",
-		"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",
-	},
-};
+/**
+ * See the registered mapping of HF model ID => Sambanova model ID here:
+ *
+ * https://huggingface.co/api/partners/sambanova/models
+ *
+ * This is a publicly available mapping.
+ *
+ * If you want to try to run inference for a new model locally before it's registered on huggingface.co,
+ * you can add it to the dictionary "HARDCODED_MODEL_ID_MAPPING" in consts.ts, for dev purposes.
+ *
+ * - If you work at Sambanova and want to update this mapping, please use the model mapping API we provide on huggingface.co
+ * - If you're a community member and want to add a new supported HF model to Sambanova, please open an issue on the present repo
+ * and we will tag Sambanova team members.
+ *
+ * Thanks!
+ */

src/providers/together.ts

@@ -1,59 +0,18 @@
-import type { ProviderMapping } from "./types";
-
 export const TOGETHER_API_BASE_URL = "https://api.together.xyz";
 
 /**
- * Same comment as in sambanova.ts
+ * See the registered mapping of HF model ID => Together model ID here:
+ *
+ * https://huggingface.co/api/partners/together/models
+ *
+ * This is a publicly available mapping.
+ *
+ * If you want to try to run inference for a new model locally before it's registered on huggingface.co,
+ * you can add it to the dictionary "HARDCODED_MODEL_ID_MAPPING" in consts.ts, for dev purposes.
+ *
+ * - If you work at Together and want to update this mapping, please use the model mapping API we provide on huggingface.co
+ * - If you're a community member and want to add a new supported HF model to Together, please open an issue on the present repo
+ * and we will tag Together team members.
+ *
+ * Thanks!
  */
-type TogetherId = string;
-
-/**
- * https://docs.together.ai/reference/models-1
- */
-export const TOGETHER_SUPPORTED_MODEL_IDS: ProviderMapping<TogetherId> = {
-	"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",
-	},
-};

No alert changes

Improved metrics

Number of package files

190

increased by1.6%

Worsened metrics

Total package byte prevSize

272136

decreased by-4.29%

Lines of code

5765

decreased by-2.6%

Number of medium supply chain risk alerts

24

increased by9.09%

Dependency changes

• Updated@huggingface/tasks@^0.15.9