lifeiteng/vall-e issues and pull requests

#80 - Prompt concatenation

Issue - State: closed - Opened by terboo54 over 1 year ago - 2 comments

#79 - AttributeError: 'CutSet' object has no attribute 'find'

Issue - State: closed - Opened by sankar-mukherjee over 1 year ago - 2 comments

#77 - Unexpected interruption during model inference

Issue - State: closed - Opened by JunZhan2000 over 1 year ago - 1 comment

#76 - A solution to the bug of AR and NAR training separately on DDP

Issue - State: closed - Opened by JunZhan2000 over 1 year ago - 3 comments

#75 - excuse me, how much times on the preprocess the audio dataset?

Issue - State: closed - Opened by friendmine over 1 year ago - 1 comment

#73 - lhotse issue

Issue - State: closed - Opened by tcmartin over 1 year ago - 1 comment

#70 - Training errors

Issue - State: closed - Opened by JianbangZ over 1 year ago - 1 comment

#68 - For chinese aishell1 egs

Pull Request - State: closed - Opened by zhaomingwork over 1 year ago - 2 comments

#67 - Sanity check with VALL-F after 17 epochs

Issue - State: closed - Opened by ghost over 1 year ago - 3 comments

#66 - There any ways this error can be fixed?

Issue - State: closed - Opened by MrCarJrProductions over 1 year ago - 1 comment

#65 - [HELP] Is there a command to use to train this on multi-gpus? If so, can you please share the commands for it?

Issue - State: closed - Opened by divineSix over 1 year ago - 2 comments

#64 - A ready-to-use docker image is created for LibriTTS

Pull Request - State: closed - Opened by aaaaaalan over 1 year ago - 1 comment

#63 - 是否支持中文数据集的训练，比如biaobei？

Issue - State: closed - Opened by zhangziliang04 over 1 year ago - 2 comments

#62 - Prefix Mode 4

Pull Request - State: closed - Opened by lifeiteng over 1 year ago

#61 - Add pypinyin g2p for Chinese

Pull Request - State: closed - Opened by zhaomingwork over 1 year ago - 5 comments

#59 - update modules

Pull Request - State: closed - Opened by lifeiteng over 1 year ago - 1 comment

#58 - After 100 epochs training, the model can synthesize natural speech on LibriTTS

Issue - State: open - Opened by dohe0342 over 1 year ago - 68 comments

#57 - any windows install document

Issue - State: closed - Opened by witchfindertr over 1 year ago - 3 comments

#56 - why only the first dim of 8

Issue - State: closed - Opened by hdmjdp over 1 year ago - 1 comment

#55 - update NAR Decoder Prefix

Pull Request - State: closed - Opened by lifeiteng over 1 year ago

#54 - Align VALL-E with the paper

Pull Request - State: closed - Opened by lifeiteng over 1 year ago

#53 - After eight hours of training, the model can synthesize natural speech!

Issue - State: closed - Opened by lifeiteng over 1 year ago - 9 comments

#52 - Support for VALL-E X

Issue - State: closed - Opened by 152334H over 1 year ago - 13 comments

#51 - The configuration of the graphics card required to train this model

Issue - State: closed - Opened by Benny0323 over 1 year ago - 1 comment

#50 - Words Mismatch during preparing LJSpeech dataset

Issue - State: closed - Opened by kin0303 over 1 year ago - 2 comments

#49 - FYI: Our WeChat group (扫码加入新一代 Speech 微信交流群)

Issue - State: open - Opened by lifeiteng over 1 year ago - 15 comments

#48 - Error during Prepare Dataset LJSpeech

Issue - State: closed - Opened by kin0303 over 1 year ago - 2 comments

#47 - grad_scale is too small, exiting

Issue - State: closed - Opened by XuanGhahahaha over 1 year ago - 6 comments

#46 - update model implementation

Pull Request - State: closed - Opened by lifeiteng over 1 year ago

#45 - Plan to train Chinese model?

Issue - State: closed - Opened by tbright17 over 1 year ago

#42 - Tune Training

Pull Request - State: closed - Opened by lifeiteng over 1 year ago - 1 comment

#41 - Add TransformerTTS and Fix Accuracy Metrics

Pull Request - State: closed - Opened by lifeiteng almost 2 years ago

#40 - Very large CUDA memory consumption in rank:0

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

#39 - goog results?

Issue - State: closed - Opened by hackerxiaobai almost 2 years ago - 1 comment

#37 - Training Status

Issue - State: closed - Opened by lifeiteng almost 2 years ago - 20 comments

#36 - [prepare data ] phonemizer :words count mismatch

Issue - State: closed - Opened by sususpe almost 2 years ago - 1 comment

#34 - Prepare Dataset not working

Issue - State: closed - Opened by do-web almost 2 years ago - 1 comment

#33 - Failure to reproduce libritts recipe

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

#32 - Icefall setup step is missing setup.py install, will cause "no module named icefall" later

Issue - State: closed - Opened by ChairGraveyard almost 2 years ago - 1 comment

#31 - Fix VALL-E Attention Mask and Tuning Optimizer

Pull Request - State: closed - Opened by lifeiteng almost 2 years ago

#30 - [Q]: How long does the "bash run.sh --stage -1 --stop-stage 3" take to finish?

Issue - State: closed - Opened by FluffyDiscord almost 2 years ago - 1 comment

#29 - The full model can be trained on a single GPU now!

Issue - State: closed - Opened by lifeiteng almost 2 years ago - 9 comments

#27 - Model not improving

Issue - State: closed - Opened by apzl almost 2 years ago - 5 comments

#26 - The relationship between AR model and the NAR model

Issue - State: closed - Opened by Maggione almost 2 years ago - 1 comment

#24 - Fix Infer errors

Pull Request - State: closed - Opened by lifeiteng almost 2 years ago

#23 - RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cpu and cuda:0! (when checking argument for argument weight in method wrapper___slow_conv2d_forward)

Issue - State: closed - Opened by Carebear80 almost 2 years ago - 3 comments

#20 - Encodec as a requirement?

Issue - State: closed - Opened by ArEnSc almost 2 years ago - 1 comment

#19 - No need for DeepSpeed

Pull Request - State: closed - Opened by lifeiteng almost 2 years ago - 1 comment

#18 - Sharing NAR Decoders and Add AdaptiveLayerNorm

Pull Request - State: closed - Opened by lifeiteng almost 2 years ago

#16 - could the dataset and data processing to be customized？

Issue - State: closed - Opened by MisakaMikoto96 almost 2 years ago - 1 comment

#15 - Discrete code vs mel-spectrogram

Issue - State: closed - Opened by yiwei0730 almost 2 years ago - 1 comment

#14 - What do the results sound like?

Issue - State: closed - Opened by Randy-H0 almost 2 years ago - 3 comments

#13 - Use Discussions First

Issue - State: closed - Opened by lifeiteng almost 2 years ago

#11 - Infer strange result

Issue - State: closed - Opened by do-web almost 2 years ago - 2 comments

#10 - Infer: Segmentation fault

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

#9 - Error: Prepare LibriTTS train/dev/test

Issue - State: closed - Opened by do-web almost 2 years ago - 1 comment

#8 - Got silence audio after inference

Issue - State: closed - Opened by jry-king almost 2 years ago - 1 comment

#7 - Infer not working

Issue - State: closed - Opened by do-web almost 2 years ago - 1 comment

#6 - tokenizer error

Issue - State: closed - Opened by do-web almost 2 years ago - 1 comment

#5 - Problems in tokenizing LibriTTS

Issue - State: closed - Opened by jry-king almost 2 years ago - 10 comments

#4 - times cost and device cost

Issue - State: closed - Opened by yiwei0730 almost 2 years ago - 3 comments

#3 - what's the difference with VALL-E and VALL-F?

Issue - State: closed - Opened by yiwei0730 almost 2 years ago - 2 comments

#2 - Implement VALL-F and VALL-E Inference

Pull Request - State: closed - Opened by lifeiteng almost 2 years ago

#1 - nano Training has been able to achieve good results

Issue - State: closed - Opened by lifeiteng almost 2 years ago - 12 comments

GitHub / lifeiteng/vall-e issues and pull requests