georgesterpu / avsr-tf1

Audio-Visual Speech Recognition using Sequence to Sequence Models
GNU General Public License v3.0
81 stars 28 forks source link

get error by running program #3

Closed wentaoxandry closed 5 years ago

georgesterpu commented 5 years ago

Hello @wentaoxandry

I am not sure if you managed to fix the problem, but it is certainly a bug on my side.

It occurs when constructing the initial state for a bimodal decoder, likely because I changed the format of the encoding_units_per_layer parameter from a tuple of ints to a tuple of tuples of ints (one for each modality) and pushed the update without running tests on this class that I rarely use.

Please feel free to submit a fix, I will also do it very soon.

wentaoxandry commented 5 years ago

Hallo @georgesterpu

Thank you for your answer, I will try my best to figure it out, but I'm new with tensorflow, I don't know if I can finish it. Please tell me if you finish this bug.

georgesterpu commented 5 years ago

@wentaoxandry Please pull and have another try now.