Why padding before passing to lstm2?

pochih / Video-Cap

🎬 Video Captioning: ICCV '15 paper implementation

48 stars 20 forks source link

Open srilekha1993 opened 4 years ago

srilekha1993 commented 4 years ago

output2, state2 = self.lstm2(tf.concat([padding, output1], 1), state2) in model.py