CPVD is a multimodal financial dataset designed specifically for candlestick pattern classification and representation learning. It combines high-quality candlestick chart visualizations with their underlying OHLCV time-series data, enabling research in domain-adapted self-supervised learning for technical pattern recognition.
The dataset comprises 1,200 labeled candlestick chart images across five distinct pattern classes… See the full description on the dataset page: https://huggingface.co/datasets/alperendemirciai/CPVD.
