People YoungJeansKR Jaeho-99 poco7250 chaelin2 ChooSeoyeon Language / IDE ML Chrome Extension Dataset 113,006 images Model OCR (CNN - LSTM - CTC Loss) Optimizer Adam Lr = 0.001 Batch size = 64 Epochs = 100 Result Train Accuracy : 45.07% Val Accuracy : 59.35% YouTube