breizhn DTLN issues - Githubissues

breizhn / DTLN

Tensorflow 2.x implementation of the DTLN real time speech denoising model. With TF-lite, ONNX and real-time audio processing support.

MIT License

567 stars 160 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Make DTLN as one model?

#80 saraphinesER closed 1 month ago
1
Why there is 384 latency?

#79 dzk9528 opened 2 months ago
0
about problem of adding win function frame

#78 yudashuixiao1 opened 4 months ago
0
Which onnx does the DTLN_norm_500h.h5 file correspond to?

#77 cpcgskill opened 5 months ago
0
error: 'tf.TensorListSetItem' op is neither a custom op nor a flex op while trying to quantize the model into int8

#76 nitolpalak opened 5 months ago
4
Integer two tflite.

#75 panhu opened 8 months ago
0
Can the calculation estimated_complex be modified?

#74 yycking closed 3 months ago
1
Multiple voices (far and near)， in this issue,how to separate

#73 mathpopo opened 11 months ago
0
Scale correction for signal reconstruction

#72 narrietal opened 1 year ago
0
The length of the audio after noise reduction is shortened

#71 1003657663 opened 1 year ago
3
Is it necessary to apply a window after fft?

#70 shawnyxf opened 1 year ago
1
Are my val_loss values valid?

#69 steven8274 closed 1 year ago
2
Using convert_weights_to_tf_lite.py did not produce the same results as the pre-training model.

#68 FrozenEelFanGirl opened 1 year ago
3
About one model

#67 panhu opened 1 year ago
1
Creating integer only models

#66 StuartIanNaylor opened 1 year ago
11
Validation loss is too high

#65 jeungmin717 opened 1 year ago
1
ValueError: Unexpected result of `train_function` (Empty logs).

#64 Morning4Star opened 1 year ago
1
How to modify DTLN_model.py files to support the latest tensorflow 2.9

#63 muyongming closed 2 years ago
0
Can DTLN be combined with AEC in real time communication?

#62 cloudvc opened 2 years ago
1
PESQ and STOI calculator

#61 klausmikealsontheking opened 2 years ago
1
Dimension mismatch error for real_time_dtln_audio.py

#60 klausmikealsontheking closed 2 years ago
1
n times of 10s data inference, the result from the first 10 seconds will be different from the next ten seconds

#59 yudashuixiao1 opened 2 years ago
0
Increasing Loss and Validation Loss

#58 hnahi closed 2 years ago
1
Question about real-time implementation using tensorflow lite

#57 seorim0 closed 2 years ago
5
Tf_lite

#56 abhinavchandel0 opened 2 years ago
9
ValueError: trying to load 500h file h5 with 10 layers but model architecture has 9 layers

#55 nikokks opened 2 years ago
5
TF_DETERMINISTIC_OPS

#54 StuartIanNaylor closed 2 years ago
2
Pesq score and data duration

#53 Fangbo0506 opened 2 years ago
1
Train stateful DTLN model

#52 chixii opened 2 years ago
6
Paper link is not working

#51 st3inum closed 2 years ago
1
TFLite Android

#50 zeynepgulhanuslu opened 2 years ago
6
Unable to test the Pre-trained model with GPU

#49 qasimali000 opened 3 years ago
1
ERROR: Could not find a version that satisfies the requirement tflite-runtime==2.1.0.post1

#48 waqasakram117 opened 3 years ago
1
Low version tf model training

#47 JuneRen opened 3 years ago
0
Issue with tflite interpreter

#46 Vanajareddy0426 opened 3 years ago
0
There is no any code with Training data preparation?

#45 shenbuguanni opened 3 years ago
2
question about B3 model in paper

#44 ghost closed 2 years ago
0
I'm training model 48K model, block_len and block_shift，encoder_size，numUnits ？

#43 ghost closed 1 year ago
3
Modify to 32K sample rate

#42 matt-audio closed 3 years ago
6
Google research

#41 StuartIanNaylor closed 3 years ago
0
PortAudioError: Error opening Stream: Invalid sample rate [PaErrorCode -9997]

#40 RXAldreezee opened 3 years ago
4
can't convert same onnx by provided python file.

#39 fmbao opened 3 years ago
1
nan when training

#38 zoeysgithub opened 3 years ago
5
Some questions about real-time denoising

#37 yolin329 opened 3 years ago
0
about traing a real time model

#36 AaZz101 closed 3 years ago
0
The data augment method used for 40h training dataset better than 500h

#35 GatsbyChen opened 3 years ago
0
The data augment method used for

#34 GatsbyChen closed 3 years ago
0
validation loss -16.83, but performance is not better than model.h5.

#33 LeeGyuHa opened 3 years ago
2
problem in running fixed size batch processing with realtime_processing_tflite.py

#32 MrGroot97 opened 3 years ago
2
The B3 model in paper

#31 GatsbyChen opened 3 years ago
0