keras - CTC alignment for video sequence recognition

Question

asked Oct 7, 2021 in Technique[技术] by 深蓝 (71.8m points)

Do v need to provide an initial alignment for the inputs during training, in addition to the actual decoded label sequence?

与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…

1.0k views

1 Answer

深蓝 · Answer 1 · 2021-10-06T19:20:37+0000

answered Oct 7, 2021 by 深蓝 (71.8m points)

Waitting for answers

与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…