@huggingface/inference 3.2.0 → 3.3.0

package/README.md

@@ -46,7 +46,12 @@ Your access token should be kept private. If you need to protect it in front-end
 
 You can send inference requests to third-party providers with the inference client.
 
-Currently, we support the following providers: [Fal.ai](https://fal.ai), [Replicate](https://replicate.com), [Together](https://together.xyz) and [Sambanova](https://sambanova.ai).
+Currently, we support the following providers:
+- [Fal.ai](https://fal.ai)
+- [Fireworks AI](https://fireworks.ai)
+- [Replicate](https://replicate.com)
+- [Sambanova](https://sambanova.ai)
+- [Together](https://together.xyz)
 
 To send requests to a third-party provider, you have to pass the `provider` parameter to the inference function. Make sure your request is authenticated with an access token.
 ```ts
@@ -64,10 +69,11 @@ When authenticated with a Hugging Face access token, the request is routed throu
 When authenticated with a third-party provider key, the request is made directly against that provider's inference API.
 
 Only a subset of models are supported when requesting third-party providers. You can check the list of supported models per pipeline tasks here:
-- [Fal.ai supported models](./src/providers/fal-ai.ts)
-- [Replicate supported models](./src/providers/replicate.ts)
-- [Sambanova supported models](./src/providers/sambanova.ts)
-- [Together supported models](./src/providers/together.ts)
+- [Fal.ai supported models](https://huggingface.co/api/partners/fal-ai/models)
+- [Fireworks AI supported models](https://huggingface.co/api/partners/fireworks-ai/models)
+- [Replicate supported models](https://huggingface.co/api/partners/replicate/models)
+- [Sambanova supported models](https://huggingface.co/api/partners/sambanova/models)
+- [Together supported models](https://huggingface.co/api/partners/together/models)
 - [HF Inference API (serverless)](https://huggingface.co/models?inference=warm&sort=trending)
 
 ❗**Important note:** To be compatible, the third-party API must adhere to the "standard" shape API we expect on HF model pages for each pipeline task type. 

package/dist/index.cjs

@@ -112,6 +112,9 @@ var SAMBANOVA_API_BASE_URL = "https://api.sambanova.ai";
 // src/providers/together.ts
 var TOGETHER_API_BASE_URL = "https://api.together.xyz";
 
+// src/providers/fireworks-ai.ts
+var FIREWORKS_AI_API_BASE_URL = "https://api.fireworks.ai/inference";
+
 // src/lib/isUrl.ts
 function isUrl(modelOrUrl) {
   return /^http(s?):/.test(modelOrUrl) || modelOrUrl.startsWith("/");
@@ -119,14 +122,22 @@ function isUrl(modelOrUrl) {
 
 // package.json
 var name = "@huggingface/inference";
-var version = "3.2.0";
+var version = "3.3.0";
 
 // src/providers/consts.ts
 var HARDCODED_MODEL_ID_MAPPING = {
   /**
    * "HF model ID" => "Model ID on Inference Provider's side"
+   *
+   * Example:
+   * "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
    */
-  // "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
+  "fal-ai": {},
+  "fireworks-ai": {},
+  "hf-inference": {},
+  replicate: {},
+  sambanova: {},
+  together: {}
 };
 
 // src/lib/getProviderModelId.ts
@@ -139,8 +150,8 @@ async function getProviderModelId(params, args, options = {}) {
     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];
+  if (HARDCODED_MODEL_ID_MAPPING[params.provider]?.[params.model]) {
+    return HARDCODED_MODEL_ID_MAPPING[params.provider][params.model];
   }
   let inferenceProviderMapping;
   if (inferenceProviderMappingCache.has(params.model)) {
@@ -291,6 +302,13 @@ function makeUrl(params) {
       }
       return baseUrl;
     }
+    case "fireworks-ai": {
+      const baseUrl = shouldProxy ? HF_HUB_INFERENCE_PROXY_TEMPLATE.replace("{{PROVIDER}}", params.provider) : FIREWORKS_AI_API_BASE_URL;
+      if (params.taskHint === "text-generation" && params.chatCompletion) {
+        return `${baseUrl}/v1/chat/completions`;
+      }
+      return baseUrl;
+    }
     default: {
       const baseUrl = HF_HUB_INFERENCE_PROXY_TEMPLATE.replaceAll("{{PROVIDER}}", "hf-inference");
       const url = params.forceTask ? `${baseUrl}/pipeline/${params.forceTask}/${params.model}` : `${baseUrl}/models/${params.model}`;
@@ -1269,7 +1287,14 @@ var HfInferenceEndpoint = class {
 };
 
 // src/types.ts
-var INFERENCE_PROVIDERS = ["fal-ai", "replicate", "sambanova", "together", "hf-inference"];
+var INFERENCE_PROVIDERS = [
+  "fal-ai",
+  "fireworks-ai",
+  "hf-inference",
+  "replicate",
+  "sambanova",
+  "together"
+];
 // Annotate the CommonJS export names for ESM import in node:
 0 && (module.exports = {
   HfInference,

package/dist/index.js

@@ -57,6 +57,9 @@ var SAMBANOVA_API_BASE_URL = "https://api.sambanova.ai";
 // src/providers/together.ts
 var TOGETHER_API_BASE_URL = "https://api.together.xyz";
 
+// src/providers/fireworks-ai.ts
+var FIREWORKS_AI_API_BASE_URL = "https://api.fireworks.ai/inference";
+
 // src/lib/isUrl.ts
 function isUrl(modelOrUrl) {
   return /^http(s?):/.test(modelOrUrl) || modelOrUrl.startsWith("/");
@@ -64,14 +67,22 @@ function isUrl(modelOrUrl) {
 
 // package.json
 var name = "@huggingface/inference";
-var version = "3.2.0";
+var version = "3.3.0";
 
 // src/providers/consts.ts
 var HARDCODED_MODEL_ID_MAPPING = {
   /**
    * "HF model ID" => "Model ID on Inference Provider's side"
+   *
+   * Example:
+   * "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
    */
-  // "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
+  "fal-ai": {},
+  "fireworks-ai": {},
+  "hf-inference": {},
+  replicate: {},
+  sambanova: {},
+  together: {}
 };
 
 // src/lib/getProviderModelId.ts
@@ -84,8 +95,8 @@ async function getProviderModelId(params, args, options = {}) {
     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];
+  if (HARDCODED_MODEL_ID_MAPPING[params.provider]?.[params.model]) {
+    return HARDCODED_MODEL_ID_MAPPING[params.provider][params.model];
   }
   let inferenceProviderMapping;
   if (inferenceProviderMappingCache.has(params.model)) {
@@ -236,6 +247,13 @@ function makeUrl(params) {
       }
       return baseUrl;
     }
+    case "fireworks-ai": {
+      const baseUrl = shouldProxy ? HF_HUB_INFERENCE_PROXY_TEMPLATE.replace("{{PROVIDER}}", params.provider) : FIREWORKS_AI_API_BASE_URL;
+      if (params.taskHint === "text-generation" && params.chatCompletion) {
+        return `${baseUrl}/v1/chat/completions`;
+      }
+      return baseUrl;
+    }
     default: {
       const baseUrl = HF_HUB_INFERENCE_PROXY_TEMPLATE.replaceAll("{{PROVIDER}}", "hf-inference");
       const url = params.forceTask ? `${baseUrl}/pipeline/${params.forceTask}/${params.model}` : `${baseUrl}/models/${params.model}`;
@@ -1214,7 +1232,14 @@ var HfInferenceEndpoint = class {
 };
 
 // src/types.ts
-var INFERENCE_PROVIDERS = ["fal-ai", "replicate", "sambanova", "together", "hf-inference"];
+var INFERENCE_PROVIDERS = [
+  "fal-ai",
+  "fireworks-ai",
+  "hf-inference",
+  "replicate",
+  "sambanova",
+  "together"
+];
 export {
   HfInference,
   HfInferenceEndpoint,

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

@@ -1 +1 @@
-{"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"}
+{"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;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

@@ -1,10 +1,13 @@
-import type { ModelId } from "../types";
+import type { InferenceProvider } from "../types";
+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.
+ *
+ * We also inject into this dictionary from tests.
  */
-export declare const HARDCODED_MODEL_ID_MAPPING: Record<ModelId, ProviderId>;
+export declare const HARDCODED_MODEL_ID_MAPPING: Record<InferenceProvider, Record<ModelId, ProviderId>>;
 export {};
 //# sourceMappingURL=consts.d.ts.map

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

@@ -1 +1 @@
-{"version":3,"file":"consts.d.ts","sourceRoot":"","sources":["../../../src/providers/consts.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,UAAU,CAAC;AAExC,KAAK,UAAU,GAAG,MAAM,CAAC;AAEzB;;;;GAIG;AACH,eAAO,MAAM,0BAA0B,EAAE,MAAM,CAAC,OAAO,EAAE,UAAU,CAKlE,CAAC"}
+{"version":3,"file":"consts.d.ts","sourceRoot":"","sources":["../../../src/providers/consts.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,iBAAiB,EAAE,MAAM,UAAU,CAAC;AAClD,OAAO,EAAE,KAAK,OAAO,EAAE,MAAM,UAAU,CAAC;AAExC,KAAK,UAAU,GAAG,MAAM,CAAC;AACzB;;;;;;GAMG;AACH,eAAO,MAAM,0BAA0B,EAAE,MAAM,CAAC,iBAAiB,EAAE,MAAM,CAAC,OAAO,EAAE,UAAU,CAAC,CAa7F,CAAC"}

package/dist/src/providers/fireworks-ai.d.ts

@@ -0,0 +1,18 @@
+export declare const FIREWORKS_AI_API_BASE_URL = "https://api.fireworks.ai/inference";
+/**
+ * See the registered mapping of HF model ID => Fireworks model ID here:
+ *
+ * https://huggingface.co/api/partners/fireworks/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 Fireworks 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 Fireworks, please open an issue on the present repo
+ * and we will tag Fireworks team members.
+ *
+ * Thanks!
+ */
+//# sourceMappingURL=fireworks-ai.d.ts.map

package/dist/src/providers/fireworks-ai.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"fireworks-ai.d.ts","sourceRoot":"","sources":["../../../src/providers/fireworks-ai.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,yBAAyB,uCAAuC,CAAC;AAE9E;;;;;;;;;;;;;;;GAeG"}

package/dist/src/types.d.ts

@@ -38,7 +38,7 @@ export interface Options {
     includeCredentials?: string | boolean;
 }
 export type InferenceTask = Exclude<PipelineType, "other">;
-export declare const INFERENCE_PROVIDERS: readonly ["fal-ai", "replicate", "sambanova", "together", "hf-inference"];
+export declare const INFERENCE_PROVIDERS: readonly ["fal-ai", "fireworks-ai", "hf-inference", "replicate", "sambanova", "together"];
 export type InferenceProvider = (typeof INFERENCE_PROVIDERS)[number];
 export interface BaseArgs {
     /**

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

@@ -1 +1 @@
-{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../src/types.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,mBAAmB,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAE5E;;GAEG;AACH,MAAM,MAAM,OAAO,GAAG,MAAM,CAAC;AAE7B,MAAM,WAAW,OAAO;IACvB;;OAEG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB;;OAEG;IACH,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB;;OAEG;IACH,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B;;OAEG;IACH,OAAO,CAAC,EAAE,OAAO,CAAC;IAElB;;OAEG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB;;OAEG;IACH,KAAK,CAAC,EAAE,OAAO,KAAK,CAAC;IACrB;;OAEG;IACH,MAAM,CAAC,EAAE,WAAW,CAAC;IAErB;;OAEG;IACH,kBAAkB,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC;CACtC;AAED,MAAM,MAAM,aAAa,GAAG,OAAO,CAAC,YAAY,EAAE,OAAO,CAAC,CAAC;AAE3D,eAAO,MAAM,mBAAmB,2EAA4E,CAAC;AAC7G,MAAM,MAAM,iBAAiB,GAAG,CAAC,OAAO,mBAAmB,CAAC,CAAC,MAAM,CAAC,CAAC;AAErE,MAAM,WAAW,QAAQ;IACxB;;;;;;OAMG;IACH,WAAW,CAAC,EAAE,MAAM,CAAC;IAErB;;;;;;;OAOG;IACH,KAAK,CAAC,EAAE,OAAO,CAAC;IAEhB;;;;OAIG;IACH,WAAW,CAAC,EAAE,MAAM,CAAC;IAErB;;;;OAIG;IACH,QAAQ,CAAC,EAAE,iBAAiB,CAAC;CAC7B;AAED,MAAM,MAAM,WAAW,GAAG,QAAQ,GACjC,CACG;IAAE,IAAI,EAAE,IAAI,GAAG,WAAW,CAAA;CAAE,GAC5B;IAAE,MAAM,EAAE,OAAO,CAAA;CAAE,GACnB;IAAE,MAAM,EAAE,MAAM,CAAA;CAAE,GAClB;IAAE,IAAI,EAAE,MAAM,CAAA;CAAE,GAChB;IAAE,SAAS,EAAE,MAAM,CAAA;CAAE,GACrB,mBAAmB,CACrB,GAAG;IACH,UAAU,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACrC,WAAW,CAAC,EAAE,MAAM,CAAC;CACrB,CAAC"}
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../src/types.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,mBAAmB,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAE5E;;GAEG;AACH,MAAM,MAAM,OAAO,GAAG,MAAM,CAAC;AAE7B,MAAM,WAAW,OAAO;IACvB;;OAEG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB;;OAEG;IACH,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB;;OAEG;IACH,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B;;OAEG;IACH,OAAO,CAAC,EAAE,OAAO,CAAC;IAElB;;OAEG;IACH,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB;;OAEG;IACH,KAAK,CAAC,EAAE,OAAO,KAAK,CAAC;IACrB;;OAEG;IACH,MAAM,CAAC,EAAE,WAAW,CAAC;IAErB;;OAEG;IACH,kBAAkB,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC;CACtC;AAED,MAAM,MAAM,aAAa,GAAG,OAAO,CAAC,YAAY,EAAE,OAAO,CAAC,CAAC;AAE3D,eAAO,MAAM,mBAAmB,2FAOtB,CAAC;AACX,MAAM,MAAM,iBAAiB,GAAG,CAAC,OAAO,mBAAmB,CAAC,CAAC,MAAM,CAAC,CAAC;AAErE,MAAM,WAAW,QAAQ;IACxB;;;;;;OAMG;IACH,WAAW,CAAC,EAAE,MAAM,CAAC;IAErB;;;;;;;OAOG;IACH,KAAK,CAAC,EAAE,OAAO,CAAC;IAEhB;;;;OAIG;IACH,WAAW,CAAC,EAAE,MAAM,CAAC;IAErB;;;;OAIG;IACH,QAAQ,CAAC,EAAE,iBAAiB,CAAC;CAC7B;AAED,MAAM,MAAM,WAAW,GAAG,QAAQ,GACjC,CACG;IAAE,IAAI,EAAE,IAAI,GAAG,WAAW,CAAA;CAAE,GAC5B;IAAE,MAAM,EAAE,OAAO,CAAA;CAAE,GACnB;IAAE,MAAM,EAAE,MAAM,CAAA;CAAE,GAClB;IAAE,IAAI,EAAE,MAAM,CAAA;CAAE,GAChB;IAAE,SAAS,EAAE,MAAM,CAAA;CAAE,GACrB,mBAAmB,CACrB,GAAG;IACH,UAAU,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACrC,WAAW,CAAC,EAAE,MAAM,CAAC;CACrB,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@huggingface/inference",
-  "version": "3.2.0",
+  "version": "3.3.0",
   "packageManager": "pnpm@8.10.5",
   "license": "MIT",
   "author": "Tim Mikeladze <tim.mikeladze@gmail.com>",

package/src/lib/getProviderModelId.ts

@@ -30,8 +30,8 @@ export async function getProviderModelId(
 		options.taskHint === "text-generation" && options.chatCompletion ? "conversational" : options.taskHint;
 
 	// A dict called HARDCODED_MODEL_ID_MAPPING takes precedence in all cases (useful for dev purposes)
-	if (HARDCODED_MODEL_ID_MAPPING[params.model]) {
-		return HARDCODED_MODEL_ID_MAPPING[params.model];
+	if (HARDCODED_MODEL_ID_MAPPING[params.provider]?.[params.model]) {
+		return HARDCODED_MODEL_ID_MAPPING[params.provider][params.model];
 	}
 
 	let inferenceProviderMapping: InferenceProviderMapping | null;

package/src/lib/makeRequestOptions.ts

@@ -3,6 +3,7 @@ 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 { FIREWORKS_AI_API_BASE_URL } from "../providers/fireworks-ai";
 import type { InferenceProvider } from "../types";
 import type { InferenceTask, Options, RequestArgs } from "../types";
 import { isUrl } from "./isUrl";
@@ -208,6 +209,15 @@ function makeUrl(params: {
 			}
 			return baseUrl;
 		}
+		case "fireworks-ai": {
+			const baseUrl = shouldProxy
+				? HF_HUB_INFERENCE_PROXY_TEMPLATE.replace("{{PROVIDER}}", params.provider)
+				: FIREWORKS_AI_API_BASE_URL;
+			if (params.taskHint === "text-generation" && params.chatCompletion) {
+				return `${baseUrl}/v1/chat/completions`;
+			}
+			return baseUrl;
+		}
 		default: {
 			const baseUrl = HF_HUB_INFERENCE_PROXY_TEMPLATE.replaceAll("{{PROVIDER}}", "hf-inference");
 			const url = params.forceTask

package/src/providers/consts.ts

@@ -1,15 +1,25 @@
-import type { ModelId } from "../types";
+import type { InferenceProvider } from "../types";
+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.
+ *
+ * We also inject into this dictionary from tests.
  */
-export const HARDCODED_MODEL_ID_MAPPING: Record<ModelId, ProviderId> = {
+export const HARDCODED_MODEL_ID_MAPPING: Record<InferenceProvider, Record<ModelId, ProviderId>> = {
 	/**
 	 * "HF model ID" => "Model ID on Inference Provider's side"
+	 *
+	 * Example:
+	 * "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
 	 */
-	// "Qwen/Qwen2.5-Coder-32B-Instruct": "Qwen2.5-Coder-32B-Instruct",
+	"fal-ai": {},
+	"fireworks-ai": {},
+	"hf-inference": {},
+	replicate: {},
+	sambanova: {},
+	together: {},
 };

package/src/providers/fireworks-ai.ts

@@ -0,0 +1,18 @@
+export const FIREWORKS_AI_API_BASE_URL = "https://api.fireworks.ai/inference";
+
+/**
+ * See the registered mapping of HF model ID => Fireworks model ID here:
+ *
+ * https://huggingface.co/api/partners/fireworks/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 Fireworks 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 Fireworks, please open an issue on the present repo
+ * and we will tag Fireworks team members.
+ *
+ * Thanks!
+ */

package/src/types.ts

@@ -44,7 +44,14 @@ export interface Options {
 
 export type InferenceTask = Exclude<PipelineType, "other">;
 
-export const INFERENCE_PROVIDERS = ["fal-ai", "replicate", "sambanova", "together", "hf-inference"] as const;
+export const INFERENCE_PROVIDERS = [
+	"fal-ai",
+	"fireworks-ai",
+	"hf-inference",
+	"replicate",
+	"sambanova",
+	"together",
+] as const;
 export type InferenceProvider = (typeof INFERENCE_PROVIDERS)[number];
 
 export interface BaseArgs {