r9y9/wavenet_vocoder issues and pull requests

#78 - Transfer learning with conditioning

Issue - State: closed - Opened by aleksas over 6 years ago - 1 comment
Labels: wontfix

#77 - [train.py] Padding value inside batched samples is wrong, especially mulaw_quantize setting.

Issue - State: closed - Opened by a3626a over 6 years ago - 3 comments
Labels: bug

#76 - CMU Arctic MuLaw samples are pitch shifted.

Issue - State: open - Opened by r9y9 over 6 years ago - 1 comment
Labels: bug

#75 - Finetuning trained model

Issue - State: closed - Opened by aleksas over 6 years ago - 3 comments

#74 - Add --preset parameter to evaluate.py.

Pull Request - State: closed - Opened by aleksas over 6 years ago - 4 comments

#73 - Remove math.sqrt(0.5) from skip connections.

Pull Request - State: closed - Opened by PetrochukM over 6 years ago - 7 comments

#72 - Random Assert Removed

Pull Request - State: closed - Opened by PetrochukM over 6 years ago

#71 - Synthesis Lengths

Pull Request - State: open - Opened by PetrochukM over 6 years ago - 2 comments

#70 - Train Frame Clip

Pull Request - State: closed - Opened by PetrochukM over 6 years ago

#69 - Make Generation Fast

Pull Request - State: closed - Opened by PetrochukM over 6 years ago - 1 comment

#68 - Dummy Silence

Issue - State: closed - Opened by PetrochukM over 6 years ago - 1 comment
Labels: wontfix

#67 - math.sqrt(0.5)

Issue - State: closed - Opened by PetrochukM over 6 years ago - 4 comments
Labels: wontfix

#66 - Sample Rate Bug

Pull Request - State: closed - Opened by PetrochukM over 6 years ago - 1 comment

#65 - Intelligible Audio

Issue - State: closed - Opened by PetrochukM over 6 years ago - 6 comments
Labels: wontfix

#64 - µ=256 while Wavenet paper used µ=255

Issue - State: closed - Opened by PetrochukM over 6 years ago - 2 comments
Labels: bug, good first issue

#63 - collate_fn in train.py doesn't apply changes in presets, causing malformed batches

Issue - State: closed - Opened by engiecat over 6 years ago - 3 comments
Labels: wontfix, windows

#62 - other dataset error?

Issue - State: closed - Opened by gudwns1215 over 6 years ago - 1 comment
Labels: wontfix

#61 - Fast Wavenet Generation Algorithm

Issue - State: closed - Opened by r9y9 over 6 years ago - 5 comments

#60 - can I change the dataset?

Issue - State: closed - Opened by gudwns1215 over 6 years ago - 1 comment

#59 - Objective evaluation of output voice quality

Issue - State: closed - Opened by danshirron over 6 years ago - 1 comment
Labels: wontfix

#58 - Fixes for PyTorch v0.4

Pull Request - State: closed - Opened by r9y9 over 6 years ago

#57 - RFC: Speech samples and pre-trained models

Issue - State: closed - Opened by r9y9 over 6 years ago - 1 comment
Labels: help wanted

#56 - implications of different input types

Issue - State: closed - Opened by danshirron over 6 years ago - 1 comment
Labels: wontfix

#55 - fix the oom problem of synthesis and fix dependecy of deep-voice3

Pull Request - State: closed - Opened by azraelkuan over 6 years ago

#54 - Support for pytorch v0.4

Issue - State: closed - Opened by r9y9 over 6 years ago - 2 comments

#53 - preprocess error while pip install -e ".[train]" OK on windows 7

Issue - State: closed - Opened by butterl over 6 years ago - 1 comment
Labels: windows

#52 - Training with Tacotron GTA mel specs

Issue - State: closed - Opened by duvtedudug over 6 years ago - 23 comments
Labels: wontfix

#51 - fix the error of oom when synthesis and remove the dependency of deepvoice3

Pull Request - State: closed - Opened by azraelkuan over 6 years ago - 5 comments

#50 - Multi-GPU Support

Issue - State: closed - Opened by ddemilla over 6 years ago - 9 comments

#49 - the problem of pre-model

Issue - State: closed - Opened by HXY-95 over 6 years ago - 4 comments
Labels: wontfix

#48 - Fix setup.py: Pytorch now supports pip install

Pull Request - State: closed - Opened by r9y9 over 6 years ago

#47 - Guiding synthesis with low-fidelity waveform?

Issue - State: closed - Opened by SvenShade over 6 years ago - 8 comments
Labels: wontfix

#46 - what is the option of the pre-trained model

Issue - State: closed - Opened by gudwns1215 over 6 years ago - 7 comments

#45 - Cannot reproduce as good audios as the demo shows

Issue - State: closed - Opened by liusongxiang over 6 years ago - 13 comments

#44 - Google releases Cloud TTS speak engine that uses wavenet

Issue - State: closed - Opened by mrsylerpowers over 6 years ago - 2 comments
Labels: wontfix

#43 - Is Implementation with Google's Cloud TPUs possible?

Issue - State: closed - Opened by mrsylerpowers over 6 years ago - 3 comments
Labels: wontfix

#42 - Speech to Speech

Issue - State: closed - Opened by PetrochukM over 6 years ago - 5 comments
Labels: question, wontfix

#41 - Training time

Issue - State: closed - Opened by PetrochukM over 6 years ago - 2 comments
Labels: question

#40 - Better Quality?

Issue - State: closed - Opened by PetrochukM over 6 years ago - 3 comments
Labels: question, wontfix

#39 - Using DeepVoice3 mels spectrogram as wavenet vocoder's input

Issue - State: closed - Opened by yishaicohe over 6 years ago - 3 comments
Labels: wontfix

#38 - Is it wrong for these code?

Issue - State: closed - Opened by luocanrao over 6 years ago - 1 comment
Labels: wontfix

#37 - Out of Memory on Synthesis

Issue - State: closed - Opened by dyelax over 6 years ago - 14 comments

#36 - Can you share with me your checkpoint file?

Issue - State: closed - Opened by einteract over 6 years ago - 1 comment

#35 - Why lws and not regular stft?

Issue - State: closed - Opened by duvtedudug over 6 years ago - 4 comments
Labels: wontfix

#34 - Is it wrong for the discretized_mix_logistic_loss?

Issue - State: closed - Opened by jiqizaisikao over 6 years ago - 3 comments
Labels: bug

#33 - It's not necessary to sum up 'log_probs' at the last dimension in discretized_mix_logistic_loss function

Issue - State: closed - Opened by naturomics over 6 years ago - 4 comments
Labels: bug

#32 - Dilation conv kernel size=3

Issue - State: closed - Opened by danshirron over 6 years ago - 2 comments
Labels: wontfix

#31 - Fixed "KeyError: 'preset'"

Pull Request - State: closed - Opened by dyelax over 6 years ago - 1 comment

#30 - Taking Tacotron2 output to wavenet vocoder

Issue - State: closed - Opened by danshirron over 6 years ago - 19 comments
Labels: wontfix

#29 - Parallel WaveNet

Issue - State: closed - Opened by kenanEkici over 6 years ago - 1 comment

#28 - About the synthesis time

Issue - State: closed - Opened by LCLL over 6 years ago - 9 comments
Labels: wontfix

#27 - Multi speaker embedding not working?!

Issue - State: closed - Opened by pfriesch over 6 years ago - 9 comments

#26 - Error in extracting melspectrogram

Issue - State: closed - Opened by Feng-long over 6 years ago - 1 comment

#25 - Thanks,i will try the mixture model again

Issue - State: closed - Opened by jiqizaisikao over 6 years ago - 1 comment
Labels: needs more info

#24 - WaveGAN vs WaveNet

Issue - State: closed - Opened by r9y9 over 6 years ago - 1 comment
Labels: wontfix

#23 - Why the discretized_mix_logistic_loss get nan？

Issue - State: closed - Opened by jiqizaisikao over 6 years ago - 2 comments
Labels: needs more info

#22 - +0.01 in audio._amp_to_db(x) ???

Issue - State: closed - Opened by duvtedudug over 6 years ago - 3 comments
Labels: bug

#21 - Added new Audio extraction + small fixes

Pull Request - State: closed - Opened by mdda over 6 years ago - 4 comments

#20 - Tkinter's PhotoImage methods and Thread Safety?

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

#19 - Good pre-trained weights anyone?

Issue - State: closed - Opened by duvtedudug over 6 years ago - 15 comments

#18 - Running out of GPU memory!

Issue - State: closed - Opened by rraallvv over 6 years ago - 12 comments
Labels: wontfix

#17 - Hi,will it convergence when the batch size is 1

Issue - State: closed - Opened by jiqizaisikao over 6 years ago - 1 comment

#16 - what's difference between forward and incremental_forward?

Issue - State: closed - Opened by zhf459 almost 7 years ago - 1 comment
Labels: question

#15 - error in synthesis

Issue - State: closed - Opened by dongmh almost 7 years ago - 8 comments

#14 - Hi,is there any proof for how discretized_mix_logistic_loss works?

Issue - State: closed - Opened by jiqizaisikao almost 7 years ago - 3 comments
Labels: wontfix

#13 - Music generation

Issue - State: closed - Opened by r9y9 almost 7 years ago - 6 comments
Labels: wontfix, discussion

#12 - Ask about Phoneme Segmentation and Phoneme Duration

Issue - State: closed - Opened by toannhu almost 7 years ago - 9 comments
Labels: question

#11 - Quick Feedback

Issue - State: closed - Opened by imdatceleste almost 7 years ago - 12 comments
Labels: wontfix

#10 - KeyError when trying synthesis.py on cmu_arctic

Issue - State: closed - Opened by ricardokleinklein almost 7 years ago - 2 comments

#9 - training

Issue - State: closed - Opened by AzamRabiee almost 7 years ago - 17 comments

#8 - arctic multi speaker generate

Issue - State: closed - Opened by azraelkuan almost 7 years ago - 4 comments

#7 - Parallel WaveNet

Issue - State: closed - Opened by DabiaoMa almost 7 years ago - 55 comments
Labels: discussion

#6 - preprocess error

Issue - State: closed - Opened by azraelkuan almost 7 years ago - 3 comments

#5 - Experimental: Mixture of logistic distributions

Pull Request - State: closed - Opened by r9y9 almost 7 years ago - 8 comments

#4 - when synthesis a wav file use synthesis.py, crash！

Issue - State: closed - Opened by neverjoe almost 7 years ago - 8 comments

#3 - Wave Noise

Issue - State: closed - Opened by weixsong almost 7 years ago - 2 comments

#2 - Add travis CI with a few cleanup

Pull Request - State: closed - Opened by r9y9 almost 7 years ago

#1 - Planned TODOs

Issue - State: closed - Opened by r9y9 almost 7 years ago - 98 comments
Labels: discussion

GitHub / r9y9/wavenet_vocoder issues and pull requests