@huggingface/tasks 0.12.22 → 0.12.24

package/src/tasks/token-classification/spec/output.json

@@ -9,7 +9,11 @@
 		"properties": {
 			"entity_group": {
 				"type": "string",
-				"description": "The predicted label for that group of tokens"
+				"description": "The predicted label for a group of one or more tokens"
+			},
+			"entity": {
+				"type": "string",
+				"description": "The predicted label for a single token"
 			},
 			"score": {
 				"type": "number",
@@ -28,6 +32,6 @@
 				"description": "The character position in the input where this group ends."
 			}
 		},
-		"required": ["label", "score"]
+		"required": ["score", "word", "start", "end"]
 	}
 }

package/src/tasks/video-text-to-text/data.ts

@@ -10,6 +10,10 @@ const taskData: TaskDataCustom = {
 			description: "A dataset of instructions and question-answer pairs about videos.",
 			id: "lmms-lab/VideoChatGPT",
 		},
+		{
+			description: "Large video understanding dataset.",
+			id: "HuggingFaceFV/finevideo",
+		},
 	],
 	demo: {
 		inputs: [
@@ -48,6 +52,10 @@ const taskData: TaskDataCustom = {
 			description: "An application to chat with a video-text-to-text model.",
 			id: "llava-hf/video-llava",
 		},
+		{
+			description: "A leaderboard for various video-text-to-text models.",
+			id: "opencompass/openvlm_video_leaderboard",
+		},
 	],
 	summary:
 		"Video-text-to-text models take in a video and a text prompt and output text. These models are also called video-language models.",

package/src/tasks/visual-question-answering/inference.ts

@@ -54,7 +54,6 @@ export interface VisualQuestionAnsweringOutputElement {
 	 * The answer to the question
 	 */
 	answer?: string;
-	label: unknown;
 	/**
 	 * The associated score / probability
 	 */

package/src/tasks/visual-question-answering/spec/output.json

@@ -16,6 +16,6 @@
 				"description": "The associated score / probability"
 			}
 		},
-		"required": ["label", "score"]
+		"required": ["score"]
 	}
 }