hiroshiyamaguchi/

dataset_123347308_nlp_sentiment_audio_text

	dataset_123347308_nlp_sentiment_audio_text.py




	


	Dataset Summary

A nlp sentiment dataset with audio text modality, stored in webdataset format.

Preprocessing: aggressive Augmentation: autoaugment

Split strategy: stratified 90 10 Sampling: curriculum

Quality filtering: moderate Labeling: self training