@huggingface/inference 2.3.3 → 2.4.0

package/README.md

@@ -211,12 +211,31 @@ await hf.tabularRegression({
   model: "scikit-learn/Fish-Weight",
   inputs: {
     data: {
-      "Height":["11.52", "12.48", "12.3778"],
-      "Length1":["23.2", "24", "23.9"],
-      "Length2":["25.4", "26.3", "26.5"],
-      "Length3":["30", "31.2", "31.1"],
-      "Species":["Bream", "Bream", "Bream"],
-      "Width":["4.02", "4.3056", "4.6961"]
+      "Height": ["11.52", "12.48", "12.3778"],
+      "Length1": ["23.2", "24", "23.9"],
+      "Length2": ["25.4", "26.3", "26.5"],
+      "Length3": ["30", "31.2", "31.1"],
+      "Species": ["Bream", "Bream", "Bream"],
+      "Width": ["4.02", "4.3056", "4.6961"]
+    },
+  },
+})
+
+await hf.tabularClassification({
+  model: "vvmnnnkv/wine-quality",
+  inputs: {
+    data: {
+      "fixed_acidity": ["7.4", "7.8", "10.3"],
+      "volatile_acidity": ["0.7", "0.88", "0.32"],
+      "citric_acid": ["0", "0", "0.45"],
+      "residual_sugar": ["1.9", "2.6", "6.4"],
+      "chlorides": ["0.076", "0.098", "0.073"],
+      "free_sulfur_dioxide": ["11", "25", "5"],
+      "total_sulfur_dioxide": ["34", "67", "13"],
+      "density": ["0.9978", "0.9968", "0.9976"],
+      "pH": ["3.51", "3.2", "3.23"],
+      "sulphates": ["0.56", "0.68", "0.82"],
+      "alcohol": ["9.4", "9.8", "12.6"]
     },
   },
 })
@@ -287,6 +306,7 @@ const { generated_text } = await gpt2.textGeneration({inputs: 'The answer to the
 ### Tabular
 
 - [x] Tabular regression
+- [x] Tabular classification
 
 ## Tree-shaking
 

package/dist/index.d.ts

@@ -448,9 +448,9 @@ export type FeatureExtractionArgs = BaseArgs & {
 	inputs: string | string[];
 };
 /**
- * Returned values are a list of floats, or a list of list of floats (depending on if you sent a string or a list of string, and if the automatic reduction, usually mean_pooling for instance was applied for you or not. This should be explained on the model's README.
+ * Returned values are a list of floats, or a list of list of floats, or a list of list of list of floats (depending on if you sent a string or a list of string, and if the automatic reduction, usually mean_pooling for instance was applied for you or not. This should be explained on the model's README.
  */
-export type FeatureExtractionOutput = (number | number[])[];
+export type FeatureExtractionOutput = (number | number[] | number[][])[];
 /**
  * This task reads some text and outputs raw float values, that are usually consumed as part of a semantic database/semantic search.
  */
@@ -875,6 +875,27 @@ export function zeroShotClassification(
 	args: ZeroShotClassificationArgs,
 	options?: Options
 ): Promise<ZeroShotClassificationOutput>;
+export type TabularClassificationArgs = BaseArgs & {
+	inputs: {
+		/**
+		 * A table of data represented as a dict of list where entries are headers and the lists are all the values, all lists must have the same size.
+		 */
+		data: Record<string, string[]>;
+	};
+};
+/**
+ * A list of predicted labels for each row
+ */
+export type TabularClassificationOutput = number[];
+/**
+ * Predicts target label for a given set of features in tabular form.
+ * Typically, you will want to train a classification model on your training data and use it with your new data of the same format.
+ * Example model: vvmnnnkv/wine-quality
+ */
+export function tabularClassification(
+	args: TabularClassificationArgs,
+	options?: Options
+): Promise<TabularClassificationOutput>;
 export type TabularRegressionArgs = BaseArgs & {
 	inputs: {
 		/**
@@ -1069,6 +1090,15 @@ export class HfInference {
 		args: Omit<ZeroShotClassificationArgs, 'accessToken'>,
 		options?: Options
 	): Promise<ZeroShotClassificationOutput>;
+	/**
+	 * Predicts target label for a given set of features in tabular form.
+	 * Typically, you will want to train a classification model on your training data and use it with your new data of the same format.
+	 * Example model: vvmnnnkv/wine-quality
+	 */
+	tabularClassification(
+		args: Omit<TabularClassificationArgs, 'accessToken'>,
+		options?: Options
+	): Promise<TabularClassificationOutput>;
 	/**
 	 * Predicts target value for a given set of features in tabular form.
 	 * Typically, you will want to train a regression model on your training data and use it with your new data of the same format.
@@ -1248,6 +1278,15 @@ export class HfInferenceEndpoint {
 		args: Omit<ZeroShotClassificationArgs, 'accessToken' | 'model'>,
 		options?: Options
 	): Promise<ZeroShotClassificationOutput>;
+	/**
+	 * Predicts target label for a given set of features in tabular form.
+	 * Typically, you will want to train a classification model on your training data and use it with your new data of the same format.
+	 * Example model: vvmnnnkv/wine-quality
+	 */
+	tabularClassification(
+		args: Omit<TabularClassificationArgs, 'accessToken' | 'model'>,
+		options?: Options
+	): Promise<TabularClassificationOutput>;
 	/**
 	 * Predicts target value for a given set of features in tabular form.
 	 * Typically, you will want to train a regression model on your training data and use it with your new data of the same format.

package/dist/index.js

@@ -41,6 +41,7 @@ __export(src_exports, {
   streamingRequest: () => streamingRequest,
   summarization: () => summarization,
   tableQuestionAnswering: () => tableQuestionAnswering,
+  tabularClassification: () => tabularClassification,
   tabularRegression: () => tabularRegression,
   textClassification: () => textClassification,
   textGeneration: () => textGeneration,
@@ -74,6 +75,7 @@ __export(tasks_exports, {
   streamingRequest: () => streamingRequest,
   summarization: () => summarization,
   tableQuestionAnswering: () => tableQuestionAnswering,
+  tabularClassification: () => tabularClassification,
   tabularRegression: () => tabularRegression,
   textClassification: () => textClassification,
   textGeneration: () => textGeneration,
@@ -459,23 +461,18 @@ async function conversational(args, options) {
 async function featureExtraction(args, options) {
   const res = await request(args, options);
   let isValidOutput = true;
-  if (Array.isArray(res)) {
-    for (const e of res) {
-      if (Array.isArray(e)) {
-        isValidOutput = e.every((x) => typeof x === "number");
-        if (!isValidOutput) {
-          break;
-        }
-      } else if (typeof e !== "number") {
-        isValidOutput = false;
-        break;
-      }
+  const isNumArrayRec = (arr, maxDepth, curDepth = 0) => {
+    if (curDepth > maxDepth)
+      return false;
+    if (arr.every((x) => Array.isArray(x))) {
+      return arr.every((x) => isNumArrayRec(x, maxDepth, curDepth + 1));
+    } else {
+      return arr.every((x) => typeof x === "number");
     }
-  } else {
-    isValidOutput = false;
-  }
+  };
+  isValidOutput = Array.isArray(res) && isNumArrayRec(res, 2, 0);
   if (!isValidOutput) {
-    throw new InferenceOutputError("Expected Array<number[] | number>");
+    throw new InferenceOutputError("Expected Array<number[][] | number[] | number>");
   }
   return res;
 }
@@ -663,6 +660,16 @@ async function tabularRegression(args, options) {
   return res;
 }
 
+// src/tasks/tabular/tabularClassification.ts
+async function tabularClassification(args, options) {
+  const res = await request(args, options);
+  const isValidOutput = Array.isArray(res) && res.every((x) => typeof x === "number");
+  if (!isValidOutput) {
+    throw new InferenceOutputError("Expected number[]");
+  }
+  return res;
+}
+
 // src/HfInference.ts
 var HfInference = class {
   accessToken;
@@ -724,6 +731,7 @@ var HfInferenceEndpoint = class {
   streamingRequest,
   summarization,
   tableQuestionAnswering,
+  tabularClassification,
   tabularRegression,
   textClassification,
   textGeneration,

package/dist/index.mjs

@@ -25,6 +25,7 @@ __export(tasks_exports, {
   streamingRequest: () => streamingRequest,
   summarization: () => summarization,
   tableQuestionAnswering: () => tableQuestionAnswering,
+  tabularClassification: () => tabularClassification,
   tabularRegression: () => tabularRegression,
   textClassification: () => textClassification,
   textGeneration: () => textGeneration,
@@ -410,23 +411,18 @@ async function conversational(args, options) {
 async function featureExtraction(args, options) {
   const res = await request(args, options);
   let isValidOutput = true;
-  if (Array.isArray(res)) {
-    for (const e of res) {
-      if (Array.isArray(e)) {
-        isValidOutput = e.every((x) => typeof x === "number");
-        if (!isValidOutput) {
-          break;
-        }
-      } else if (typeof e !== "number") {
-        isValidOutput = false;
-        break;
-      }
+  const isNumArrayRec = (arr, maxDepth, curDepth = 0) => {
+    if (curDepth > maxDepth)
+      return false;
+    if (arr.every((x) => Array.isArray(x))) {
+      return arr.every((x) => isNumArrayRec(x, maxDepth, curDepth + 1));
+    } else {
+      return arr.every((x) => typeof x === "number");
     }
-  } else {
-    isValidOutput = false;
-  }
+  };
+  isValidOutput = Array.isArray(res) && isNumArrayRec(res, 2, 0);
   if (!isValidOutput) {
-    throw new InferenceOutputError("Expected Array<number[] | number>");
+    throw new InferenceOutputError("Expected Array<number[][] | number[] | number>");
   }
   return res;
 }
@@ -614,6 +610,16 @@ async function tabularRegression(args, options) {
   return res;
 }
 
+// src/tasks/tabular/tabularClassification.ts
+async function tabularClassification(args, options) {
+  const res = await request(args, options);
+  const isValidOutput = Array.isArray(res) && res.every((x) => typeof x === "number");
+  if (!isValidOutput) {
+    throw new InferenceOutputError("Expected number[]");
+  }
+  return res;
+}
+
 // src/HfInference.ts
 var HfInference = class {
   accessToken;
@@ -674,6 +680,7 @@ export {
   streamingRequest,
   summarization,
   tableQuestionAnswering,
+  tabularClassification,
   tabularRegression,
   textClassification,
   textGeneration,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@huggingface/inference",
-  "version": "2.3.3",
+  "version": "2.4.0",
   "packageManager": "pnpm@8.3.1",
   "license": "MIT",
   "author": "Tim Mikeladze <tim.mikeladze@gmail.com>",

package/src/tasks/index.ts

@@ -36,3 +36,4 @@ export * from "./multimodal/visualQuestionAnswering";
 
 // Tabular tasks
 export * from "./tabular/tabularRegression";
+export * from "./tabular/tabularClassification";

package/src/tasks/nlp/featureExtraction.ts

@@ -13,9 +13,9 @@ export type FeatureExtractionArgs = BaseArgs & {
 };
 
 /**
- * Returned values are a list of floats, or a list of list of floats (depending on if you sent a string or a list of string, and if the automatic reduction, usually mean_pooling for instance was applied for you or not. This should be explained on the model's README.
+ * Returned values are a list of floats, or a list of list of floats, or a list of list of list of floats (depending on if you sent a string or a list of string, and if the automatic reduction, usually mean_pooling for instance was applied for you or not. This should be explained on the model's README.
  */
-export type FeatureExtractionOutput = (number | number[])[];
+export type FeatureExtractionOutput = (number | number[] | number[][])[];
 
 /**
  * This task reads some text and outputs raw float values, that are usually consumed as part of a semantic database/semantic search.
@@ -26,26 +26,20 @@ export async function featureExtraction(
 ): Promise<FeatureExtractionOutput> {
 	const res = await request<FeatureExtractionOutput>(args, options);
 	let isValidOutput = true;
-	// Check if output is an array
-	if (Array.isArray(res)) {
-		for (const e of res) {
-			// Check if output is an array of arrays or numbers
-			if (Array.isArray(e)) {
-				// if all elements are numbers, continue
-				isValidOutput = e.every((x) => typeof x === "number");
-				if (!isValidOutput) {
-					break;
-				}
-			} else if (typeof e !== "number") {
-				isValidOutput = false;
-				break;
-			}
+
+	const isNumArrayRec = (arr: unknown[], maxDepth: number, curDepth = 0): boolean => {
+		if (curDepth > maxDepth) return false;
+		if (arr.every((x) => Array.isArray(x))) {
+			return arr.every((x) => isNumArrayRec(x as unknown[], maxDepth, curDepth + 1));
+		} else {
+			return arr.every((x) => typeof x === "number");
 		}
-	} else {
-		isValidOutput = false;
-	}
+	};
+
+	isValidOutput = Array.isArray(res) && isNumArrayRec(res, 2, 0);
+
 	if (!isValidOutput) {
-		throw new InferenceOutputError("Expected Array<number[] | number>");
+		throw new InferenceOutputError("Expected Array<number[][] | number[] | number>");
 	}
 	return res;
 }

package/src/tasks/tabular/tabularClassification.ts

@@ -0,0 +1,34 @@
+import { InferenceOutputError } from "../../lib/InferenceOutputError";
+import type { BaseArgs, Options } from "../../types";
+import { request } from "../custom/request";
+
+export type TabularClassificationArgs = BaseArgs & {
+	inputs: {
+		/**
+		 * A table of data represented as a dict of list where entries are headers and the lists are all the values, all lists must have the same size.
+		 */
+		data: Record<string, string[]>;
+	};
+};
+
+/**
+ * A list of predicted labels for each row
+ */
+export type TabularClassificationOutput = number[];
+
+/**
+ * Predicts target label for a given set of features in tabular form.
+ * Typically, you will want to train a classification model on your training data and use it with your new data of the same format.
+ * Example model: vvmnnnkv/wine-quality
+ */
+export async function tabularClassification(
+	args: TabularClassificationArgs,
+	options?: Options
+): Promise<TabularClassificationOutput> {
+	const res = await request<TabularClassificationOutput>(args, options);
+	const isValidOutput = Array.isArray(res) && res.every((x) => typeof x === "number");
+	if (!isValidOutput) {
+		throw new InferenceOutputError("Expected number[]");
+	}
+	return res;
+}