{
    "dataset": "emotion_classification",
    "task_type": "dialog_emotion_recognition_multimodal",
    "total_target_emotions": 7,
    "valid_target_emotions": [
        "positive_joy",
        "positive_surprise",
        "negative_anger",
        "negative_surprise",
        "negative_disgust",
        "negative_fear",
        "negative_sadness"
    ],
    "questions_per_emotion": 100,
    "total_questions_per_combination": 700,
    "emotion_stats": {
        "positive_joy": 100,
        "positive_surprise": 100,
        "negative_anger": 100,
        "negative_surprise": 100,
        "negative_disgust": 100,
        "negative_fear": 100,
        "negative_sadness": 100
    },
    "combinations": [
        "audio_vision",
        "vision_audio",
        "vision_text",
        "text_vision",
        "text_audio",
        "audio_text"
    ],
    "sampling_strategy": "1_correct + 1_neutral + 2_opposite_category",
    "face_media_type": "mp4_videos",
    "audio_media_type": "wav_dialog_files",
    "task_description": "Cross-modal emotion recognition from dialog and conversation data"
}