search

JusperLee / CTCNet

An Audio-Visual Speech Separation Model Inspired by Cortico-Thalamo-Cortical Circuits
Apache License 2.0
68 stars 16 forks source link

ctcnet,py里的def fuse #13

Open Fenglingling9420 opened 6 months ago

Fenglingling9420 commented 6 months ago

'VideoBlock' object has no attribute 'get_block_block'. Did you mean: 'get_concat_block'?

JusperLee commented 6 months ago

get_video_block

Fenglingling9420 commented 6 months ago

训练的时候loss是负的是正常现象吗

JusperLee commented 6 months ago

是正常的。优化方向使朝向最小

Fenglingling9420 commented 6 months ago

hello，n_src设置为1得到的参数就是只能分离一个人的语音吗

JusperLee commented 3 months ago

你可以迭代的分离不同的人

Fenglingling9420 commented 3 months ago

为什么有时候我在训练的时候底下的日志是false 32000，有时候却是正常的 False 32000 ['/root/autodl-tmp/CTCNet-main/Datasets/LRS2/audio/tr/mix/5535496873950688380_00015_0.3814_607858775758877234500009-0.3814.wav', 32000] False 32000 ['/root/autodl-tmp/CTCNet-main/Datasets/LRS2/audio/tr/mix/6132634767048085803_00049_0.70108_620351461233397449100033-0.70108.wav', 32000] False 32000 ['/root/autodl-tmp/CTCNet-main/Datasets/LRS2/audio/tr/mix/5964019075471165468_00010_4.5628_568742146362192590700008-4.5628.wav', 32000] False 32000 ['/root/autodl-tmp/CTCNet-main/Datasets/LRS2/audio/tr/mix/5963261443240215512_00005_4.6592_598360799181154830800008-4.6592.wav', 32000]