@huggingface/inference 3.1.0 → 3.1.2

package/README.md

@@ -1,4 +1,4 @@
-# 🤗 Hugging Face Inference Endpoints
+# 🤗 Hugging Face Inference
 
 A Typescript powered wrapper for the Hugging Face 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.

package/dist/index.cjs

@@ -54,6 +54,7 @@ __export(src_exports, {
   textGenerationStream: () => textGenerationStream,
   textToImage: () => textToImage,
   textToSpeech: () => textToSpeech,
+  textToVideo: () => textToVideo,
   tokenClassification: () => tokenClassification,
   translation: () => translation,
   visualQuestionAnswering: () => visualQuestionAnswering,
@@ -91,6 +92,7 @@ __export(tasks_exports, {
   textGenerationStream: () => textGenerationStream,
   textToImage: () => textToImage,
   textToSpeech: () => textToSpeech,
+  textToVideo: () => textToVideo,
   tokenClassification: () => tokenClassification,
   translation: () => translation,
   visualQuestionAnswering: () => visualQuestionAnswering,
@@ -911,6 +913,36 @@ async function zeroShotImageClassification(args, options) {
   return res;
 }
 
+// src/tasks/cv/textToVideo.ts
+var SUPPORTED_PROVIDERS = ["fal-ai", "replicate"];
+async function textToVideo(args, options) {
+  if (!args.provider || !typedInclude(SUPPORTED_PROVIDERS, args.provider)) {
+    throw new Error(
+      `textToVideo inference is only supported for the following providers: ${SUPPORTED_PROVIDERS.join(", ")}`
+    );
+  }
+  const payload = args.provider === "fal-ai" || args.provider === "replicate" ? { ...omit(args, ["inputs", "parameters"]), ...args.parameters, prompt: args.inputs } : args;
+  const res = await request(payload, {
+    ...options,
+    taskHint: "text-to-video"
+  });
+  if (args.provider === "fal-ai") {
+    const isValidOutput = typeof res === "object" && !!res && "video" in res && typeof res.video === "object" && !!res.video && "url" in res.video && typeof res.video.url === "string" && isUrl(res.video.url);
+    if (!isValidOutput) {
+      throw new InferenceOutputError("Expected { video: { url: string } }");
+    }
+    const urlResponse = await fetch(res.video.url);
+    return await urlResponse.blob();
+  } else {
+    const isValidOutput = typeof res === "object" && !!res && "output" in res && typeof res.output === "string" && isUrl(res.output);
+    if (!isValidOutput) {
+      throw new InferenceOutputError("Expected { output: string }");
+    }
+    const urlResponse = await fetch(res.output);
+    return await urlResponse.blob();
+  }
+}
+
 // src/lib/getDefaultTask.ts
 var taskCache = /* @__PURE__ */ new Map();
 var CACHE_DURATION = 10 * 60 * 1e3;
@@ -1338,6 +1370,7 @@ var INFERENCE_PROVIDERS = ["fal-ai", "replicate", "sambanova", "together", "hf-i
   textGenerationStream,
   textToImage,
   textToSpeech,
+  textToVideo,
   tokenClassification,
   translation,
   visualQuestionAnswering,

package/dist/index.js

@@ -33,6 +33,7 @@ __export(tasks_exports, {
   textGenerationStream: () => textGenerationStream,
   textToImage: () => textToImage,
   textToSpeech: () => textToSpeech,
+  textToVideo: () => textToVideo,
   tokenClassification: () => tokenClassification,
   translation: () => translation,
   visualQuestionAnswering: () => visualQuestionAnswering,
@@ -853,6 +854,36 @@ async function zeroShotImageClassification(args, options) {
   return res;
 }
 
+// src/tasks/cv/textToVideo.ts
+var SUPPORTED_PROVIDERS = ["fal-ai", "replicate"];
+async function textToVideo(args, options) {
+  if (!args.provider || !typedInclude(SUPPORTED_PROVIDERS, args.provider)) {
+    throw new Error(
+      `textToVideo inference is only supported for the following providers: ${SUPPORTED_PROVIDERS.join(", ")}`
+    );
+  }
+  const payload = args.provider === "fal-ai" || args.provider === "replicate" ? { ...omit(args, ["inputs", "parameters"]), ...args.parameters, prompt: args.inputs } : args;
+  const res = await request(payload, {
+    ...options,
+    taskHint: "text-to-video"
+  });
+  if (args.provider === "fal-ai") {
+    const isValidOutput = typeof res === "object" && !!res && "video" in res && typeof res.video === "object" && !!res.video && "url" in res.video && typeof res.video.url === "string" && isUrl(res.video.url);
+    if (!isValidOutput) {
+      throw new InferenceOutputError("Expected { video: { url: string } }");
+    }
+    const urlResponse = await fetch(res.video.url);
+    return await urlResponse.blob();
+  } else {
+    const isValidOutput = typeof res === "object" && !!res && "output" in res && typeof res.output === "string" && isUrl(res.output);
+    if (!isValidOutput) {
+      throw new InferenceOutputError("Expected { output: string }");
+    }
+    const urlResponse = await fetch(res.output);
+    return await urlResponse.blob();
+  }
+}
+
 // src/lib/getDefaultTask.ts
 var taskCache = /* @__PURE__ */ new Map();
 var CACHE_DURATION = 10 * 60 * 1e3;
@@ -1279,6 +1310,7 @@ export {
   textGenerationStream,
   textToImage,
   textToSpeech,
+  textToVideo,
   tokenClassification,
   translation,
   visualQuestionAnswering,

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

@@ -11,6 +11,7 @@ export * from "./cv/objectDetection";
 export * from "./cv/textToImage";
 export * from "./cv/imageToImage";
 export * from "./cv/zeroShotImageClassification";
+export * from "./cv/textToVideo";
 export * from "./nlp/featureExtraction";
 export * from "./nlp/fillMask";
 export * from "./nlp/questionAnswering";

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

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/tasks/index.ts"],"names":[],"mappings":"AACA,cAAc,kBAAkB,CAAC;AACjC,cAAc,2BAA2B,CAAC;AAG1C,cAAc,6BAA6B,CAAC;AAC5C,cAAc,oCAAoC,CAAC;AACnD,cAAc,sBAAsB,CAAC;AACrC,cAAc,sBAAsB,CAAC;AAGrC,cAAc,0BAA0B,CAAC;AACzC,cAAc,wBAAwB,CAAC;AACvC,cAAc,kBAAkB,CAAC;AACjC,cAAc,sBAAsB,CAAC;AACrC,cAAc,kBAAkB,CAAC;AACjC,cAAc,mBAAmB,CAAC;AAClC,cAAc,kCAAkC,CAAC;AAGjD,cAAc,yBAAyB,CAAC;AACxC,cAAc,gBAAgB,CAAC;AAC/B,cAAc,yBAAyB,CAAC;AACxC,cAAc,0BAA0B,CAAC;AACzC,cAAc,qBAAqB,CAAC;AACpC,cAAc,8BAA8B,CAAC;AAC7C,cAAc,0BAA0B,CAAC;AACzC,cAAc,sBAAsB,CAAC;AACrC,cAAc,4BAA4B,CAAC;AAC3C,cAAc,2BAA2B,CAAC;AAC1C,cAAc,mBAAmB,CAAC;AAClC,cAAc,8BAA8B,CAAC;AAC7C,cAAc,sBAAsB,CAAC;AACrC,cAAc,4BAA4B,CAAC;AAG3C,cAAc,wCAAwC,CAAC;AACvD,cAAc,sCAAsC,CAAC;AAGrD,cAAc,6BAA6B,CAAC;AAC5C,cAAc,iCAAiC,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/tasks/index.ts"],"names":[],"mappings":"AACA,cAAc,kBAAkB,CAAC;AACjC,cAAc,2BAA2B,CAAC;AAG1C,cAAc,6BAA6B,CAAC;AAC5C,cAAc,oCAAoC,CAAC;AACnD,cAAc,sBAAsB,CAAC;AACrC,cAAc,sBAAsB,CAAC;AAGrC,cAAc,0BAA0B,CAAC;AACzC,cAAc,wBAAwB,CAAC;AACvC,cAAc,kBAAkB,CAAC;AACjC,cAAc,sBAAsB,CAAC;AACrC,cAAc,kBAAkB,CAAC;AACjC,cAAc,mBAAmB,CAAC;AAClC,cAAc,kCAAkC,CAAC;AACjD,cAAc,kBAAkB,CAAC;AAGjC,cAAc,yBAAyB,CAAC;AACxC,cAAc,gBAAgB,CAAC;AAC/B,cAAc,yBAAyB,CAAC;AACxC,cAAc,0BAA0B,CAAC;AACzC,cAAc,qBAAqB,CAAC;AACpC,cAAc,8BAA8B,CAAC;AAC7C,cAAc,0BAA0B,CAAC;AACzC,cAAc,sBAAsB,CAAC;AACrC,cAAc,4BAA4B,CAAC;AAC3C,cAAc,2BAA2B,CAAC;AAC1C,cAAc,mBAAmB,CAAC;AAClC,cAAc,8BAA8B,CAAC;AAC7C,cAAc,sBAAsB,CAAC;AACrC,cAAc,4BAA4B,CAAC;AAG3C,cAAc,wCAAwC,CAAC;AACvD,cAAc,sCAAsC,CAAC;AAGrD,cAAc,6BAA6B,CAAC;AAC5C,cAAc,iCAAiC,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@huggingface/inference",
-  "version": "3.1.0",
+  "version": "3.1.2",
   "packageManager": "pnpm@8.10.5",
   "license": "MIT",
   "author": "Tim Mikeladze <tim.mikeladze@gmail.com>",
@@ -39,7 +39,7 @@
   },
   "type": "module",
   "dependencies": {
-    "@huggingface/tasks": "^0.14.0"
+    "@huggingface/tasks": "^0.15.0"
   },
   "devDependencies": {
     "@types/node": "18.13.0"

package/src/tasks/index.ts

@@ -16,6 +16,7 @@ export * from "./cv/objectDetection";
 export * from "./cv/textToImage";
 export * from "./cv/imageToImage";
 export * from "./cv/zeroShotImageClassification";
+export * from "./cv/textToVideo";
 
 // Natural Language Processing tasks
 export * from "./nlp/featureExtraction";