Upload dataset

README.md

@@ -1,16 +1,55 @@
 ---
-license: cc-by-nc-sa-4.0
 language:
 - en
-pretty_name: ' Video2Reaction '
+license: cc-by-nc-sa-4.0
 size_categories:
 - 1B<n<10B
 task_categories:
 - other
+pretty_name: ' Video2Reaction '
 tags:
 - video
 - audio
 - text
+dataset_info:
+  features:
+  - name: video_id
+    dtype: string
+  - name: reaction_dominant
+    dtype: string
+  - name: movie_genre
+    dtype: string
+  - name: clip_description_embedding
+    dtype: string
+  - name: visual_feature
+    dtype: string
+  - name: audio_acoustic_feature
+    dtype: string
+  - name: audio_semantic_feature
+    dtype: string
+  - name: reaction_distribution
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 1873374
+    num_examples: 7243
+  - name: val
+    num_bytes: 267702
+    num_examples: 1035
+  - name: test
+    num_bytes: 535383
+    num_examples: 2070
+  download_size: 1131847
+  dataset_size: 2676459
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+  - split: val
+    path: data/val-*
+  - split: test
+    path: data/test-*
 ---
 
 In this paper, we introduce **Video2Reaction**, a large-scale dataset consisting of over 10,000 movie clips sourced from the licensed MovieClips YouTube channel. Each video is paired with audience comments, allowing for a precise mapping between visual content and the emotional reactions it induces. Unlike perceived emotions, which are typically modeled as unimodal (single-label), induced emotions can be either unimodal or split across multiple emotions. This distinction makes it more important to learn the distribution of reactions, rather than simply predicting single or multi-class labels. To address this, we frame audience emotion recognition as a **label distribution learning** (LDL) problem. Rather than classifying a single dominant reaction, we model the distribution of emotional responses from the population for each video, enabling us to capture the diverse and nuanced nature of audience reactions.

data/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c84f1f285c1b04896be93d9b95886c00310debcd8e8545d52b8a5628ca3c7dd
+size 228744

data/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:117f1c16c0f7c79c22c1629f317d2882250d149501241d5ca388437a6d231745
+size 785961

data/val-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:840b64d517d507c5328fd729fd74f4db370265e7dad9edffbe9d1fb02dd4b38d
+size 117142