clovaai/voxceleb_trainer issues and pull requests

#84 - View size is not compatible with input tensor's size

Issue - State: closed - Opened by surasakBoonkla almost 4 years ago - 4 comments

#83 - Insufficient shared memory

Issue - State: closed - Opened by surasakBoonkla almost 4 years ago - 2 comments

#82 - Problems with VGGVox's result

Issue - State: closed - Opened by huyvl almost 4 years ago - 4 comments

#81 - Distributed and mixed precision training

Pull Request - State: closed - Opened by joonson almost 4 years ago

#80 - Speed up Speed down

Issue - State: closed - Opened by hangxiu almost 4 years ago - 1 comment

#79 - THIS ISSUE WAS DELETED AND BLOCKED

Issue - State: closed - Opened by vuongvantinh about 4 years ago - 3 comments

#78 - GPU utilization is 0

Issue - State: closed - Opened by hangxiu about 4 years ago - 1 comment

#77 - train_list not found

Issue - State: closed - Opened by rohitpaturi about 4 years ago

#76 - Hyperparameters about ResNetSE34V2

Issue - State: closed - Opened by whitegon about 4 years ago - 1 comment

#75 - Assertion Error when trainning VGGVox

Issue - State: closed - Opened by huyvl about 4 years ago - 1 comment

#74 - using distribute the train porcess can not exit when training is end

Issue - State: closed - Opened by zh794390558 about 4 years ago - 2 comments

#73 - cleaned trial pairs of vox1

Issue - State: closed - Opened by Shane-pe about 4 years ago - 2 comments

#72 - New to this field

Issue - State: closed - Opened by sbkim052 about 4 years ago - 1 comment

#71 - Reproducing ResNetSE34V2

Issue - State: closed - Opened by skubicius about 4 years ago - 1 comment

#70 - How to generate trial_pair_Verification.txt file if I would like to use own datasets.

Issue - State: closed - Opened by ASR2020Guru about 4 years ago - 4 comments

#69 - how much improve by PreEmphasis?

Issue - State: closed - Opened by zh794390558 about 4 years ago - 1 comment

#68 - In eval phase, is it correct to use L2 distance even for Angular xx losses?

Issue - State: closed - Opened by w4-jonghoon about 4 years ago - 2 comments

#67 - Training speed difference between hdd and ssd

Issue - State: closed - Opened by gustmd0121 about 4 years ago - 1 comment

#66 - Replace argument encoder to encoder_type

Pull Request - State: closed - Opened by ngoanpv about 4 years ago - 1 comment

#65 - OSError: [Errno 5] Input/output error

Issue - State: closed - Opened by dendi123 about 4 years ago - 2 comments

#64 - Rename nSpeakers to nClasses in readme

Pull Request - State: closed - Opened by dvisockas about 4 years ago

#63 - About the data augment

Issue - State: closed - Opened by wuqiangch about 4 years ago - 4 comments

#62 - Training details for the new best model

Issue - State: closed - Opened by hwidong-na about 4 years ago - 2 comments

#61 - Add distributed training

Issue - State: closed - Opened by joonson about 4 years ago - 11 comments

#60 - Regarding Cross Attentive Pooling

Issue - State: closed - Opened by 009deep about 4 years ago - 2 comments

#59 - Add data augmentation

Issue - State: closed - Opened by joonson about 4 years ago - 2 comments

#58 - ResNetSE34 result question

Issue - State: closed - Opened by yy835055664 about 4 years ago - 3 comments

#57 - SEBasicBlock operation order

Issue - State: closed - Opened by cirograciapons about 4 years ago - 1 comment

#56 - ValueError: source code string cannot contain null bytes

Issue - State: closed - Opened by dendi123 about 4 years ago - 2 comments

#55 - Discussions for training / VoxSRC

Issue - State: closed - Opened by joonson about 4 years ago - 33 comments

#54 - Is the thinResNet34 with angular-prototypical the state-of-the-art neural network you know?

Issue - State: closed - Opened by zeek-han about 4 years ago - 4 comments

#53 - Problems in converting voxceleb

Issue - State: closed - Opened by huyvl about 4 years ago - 3 comments

#52 - A problem about prototypical loss.

Issue - State: closed - Opened by XinhaoMei about 4 years ago - 2 comments

#51 - Fast ResNet-34?

Issue - State: closed - Opened by XinhaoMei over 4 years ago - 2 comments

#50 - don't we need tuneThreshold.py?

Issue - State: closed - Opened by zeek-han over 4 years ago - 1 comment

#49 - my score file has ONLY NEGATIVE score....

Issue - State: closed - Opened by zeek-han over 4 years ago - 5 comments

#48 - Regarding SAP implemantaion

Issue - State: closed - Opened by 009deep over 4 years ago - 1 comment

#47 - Why using different input features

Issue - State: closed - Opened by XinhaoMei over 4 years ago - 1 comment

#46 - Does nPerSpeaker in the code equal to M in the paper?

Issue - State: closed - Opened by XinhaoMei over 4 years ago - 1 comment

#45 - I need not only test accuracy but test loss also...

Issue - State: closed - Opened by zeek-han over 4 years ago - 4 comments

#44 - nSpeaker affects M?

Issue - State: closed - Opened by KurtAhn over 4 years ago - 1 comment

#43 - Generate embeddings of a long audio file

Issue - State: closed - Opened by Joepetey over 4 years ago - 1 comment

#42 - why the speaker model can not train by multi gpus?

Issue - State: closed - Opened by zh794390558 over 4 years ago - 1 comment

#41 - how do I train angular-prototypical loss for M=2?

Issue - State: closed - Opened by zeek-han over 4 years ago - 4 comments

#40 - Experimental results

Issue - State: closed - Opened by yy835055664 over 4 years ago - 4 comments

#39 - Questions about the test list

Issue - State: closed - Opened by yy835055664 over 4 years ago - 1 comment

#38 - Overfitting using VGG?

Issue - State: closed - Opened by XinhaoMei over 4 years ago - 3 comments

#37 - train_list file not found

Issue - State: closed - Opened by lyc1993 over 4 years ago - 1 comment

#36 - Pairwise for 10 features when evaluation

Issue - State: closed - Opened by jlian2 over 4 years ago - 1 comment

#35 - Is `test_list` flag intended to point to "the test list for VoxCeleb1"?

Issue - State: closed - Opened by w4-jonghoon over 4 years ago - 3 comments

#34 - Angular Prototypical and # of classes

Issue - State: closed - Opened by jlian2 over 4 years ago - 1 comment

#33 - Any suggestions on reproduce?

Issue - State: closed - Opened by jlian2 over 4 years ago - 3 comments

#32 - Comparing to your previous paper(VoxCeleb2)

Issue - State: closed - Opened by jlian2 over 4 years ago - 2 comments

#31 - Regarding 500 training epochs

Issue - State: closed - Opened by jlian2 over 4 years ago - 1 comment

#30 - MelSpectrum and stft

Issue - State: closed - Opened by jlian2 over 4 years ago - 1 comment

#29 - Convergence

Issue - State: closed - Opened by jlian2 over 4 years ago - 6 comments

#28 - Cannot Find Dev/Test Lists

Issue - State: closed - Opened by mmwebster over 4 years ago - 1 comment

#27 - Merge pull request #1 from clovaai/master

Pull Request - State: closed - Opened by matln over 4 years ago

#26 - Multi gpu cpu support

Pull Request - State: closed - Opened by 009deep over 4 years ago - 9 comments

#25 - issues with using triplet as loss function to train the model

Issue - State: closed - Opened by pengcheng-tech over 4 years ago - 2 comments

#24 - Can't replicate the results

Issue - State: closed - Opened by mussakhojayeva over 4 years ago - 7 comments

#23 - Incorrect argument in readme.

Issue - State: closed - Opened by Derpimort over 4 years ago - 1 comment

#22 - Consistent metric for training and testing

Issue - State: closed - Opened by hwidong-na over 4 years ago - 2 comments

#21 - about AMsoftmax

Issue - State: closed - Opened by DYJ1111 over 4 years ago - 7 comments

#20 - Pretrained models for GE2E

Issue - State: closed - Opened by Edresson over 4 years ago - 4 comments

#19 - add multiprocessing to conversion task

Pull Request - State: closed - Opened by rvirgilli over 4 years ago

#18 - About nSpeakers setting

Issue - State: closed - Opened by Erunning over 4 years ago - 1 comment

#17 - Pretrained models for AM/AMS-Softmax

Issue - State: closed - Opened by mussakhojayeva over 4 years ago - 5 comments

#16 - Are their any pretrained models for MFCC/CQCC

Issue - State: closed - Opened by jlian2 over 4 years ago - 1 comment

#15 - Regarding negative similarity score

Issue - State: closed - Opened by 009deep over 4 years ago - 3 comments

#14 - A wrong argument in trainSpeakerNet.py

Issue - State: closed - Opened by XinhaoMei over 4 years ago - 1 comment

#13 - Implementación

Issue - State: closed - Opened by franyerdev over 4 years ago

#12 - Fast ResNet-34 specifications

Issue - State: closed - Opened by hyperia-steven over 4 years ago - 2 comments

GitHub / clovaai/voxceleb_trainer issues and pull requests