search

rizkiarm / LipNet

Keras implementation of 'LipNet: End-to-End Sentence-level Lipreading'
MIT License
638 stars 226 forks source link

Implement V2P net #82

Open LJQCN101 opened 5 years ago

LJQCN101 commented 5 years ago

Implement V2P network structure according to paper "LARGE-SCALE VISUAL SPEECH RECOGNITION" (https://arxiv.org/abs/1807.05162)

Requires keras-contrib. (https://github.com/keras-team/keras-contrib)

akouminov commented 4 years ago

Hi! I tried this out and had a little trouble running it,

I was wondering if you could help me!

Here is the screenshot of the error

image

LJQCN101 commented 4 years ago

Hi! I tried this out and had a little trouble running it,

I was wondering if you could help me!

Here is the screenshot of the error

image

Hi, according to the paper, input image size should be 128x128. Try img_w=128, img_h=128. You may want to resize the image first.