facebookarchive / loop

A method to generate speech across multiple speakers

facebookarchive/loop Issues

Block on preprocessing
Updated 5 years ago
hello world text
Closed 5 years ago
TBPTTIter.split_length() error
Updated 5 years ago1
Batch
Updated 5 years ago
Strange fail on "The quick brown fox jumps over the lazy dog."
Updated 5 years ago
Look like it fails on '!' character.
Updated 5 years ago
ERROR: Failed to find norm file.
Closed 5 years ago7
Understanding feat tensor dimensions
Updated 5 years ago1
How this repo compared to Merlin?
Updated 5 years ago
ImportError: No module named torch
Closed 5 years ago
bash scripts/download_tools.sh failed on Mac OS
Closed 5 years ago
Parameters for dataset in the wild
Updated 5 years ago12
Train VCTK dataset for all speakers
Updated 5 years ago
Error when 'make' HTK-3.4.1 and hts_core.
Closed 5 years ago
Error running train.py
Updated 5 years ago1
No matching distribution found for phonemizer (from -r scripts/requirements.txt (line 5))
Closed 6 years ago1
Any clue? Missing norm_info_mgc_lf0_vuv_bap_63_MVN.dat while prepossessing Lj Speech Dataset.
Updated 6 years ago7
Issue running install_tts.py to preprocess data
Closed 6 years ago1
Main Readme wav files are missing and first instruction doesn't work
Updated 6 years ago1
Using pre-trained model for new speaker?
Updated 6 years ago
Trump voice
Closed 6 years ago1
Out of memory in validation step
Closed 6 years ago1
Error while generating sample speech from pretrained model
Closed 6 years ago2
How much minutes of audio datasets to train for a single speaker using blizzard model?
Closed 7 years ago47
Recreating blizzard baseline
Updated 6 years ago
accuracy of the model by using vctk
Updated 6 years ago
Training Error. Epoch 5.
Closed 6 years ago3
visualize flag not working with visdom
Closed 6 years ago1
What does a good loss curve look like?
Closed 6 years ago2
How much hours or minutes of training data do we need to make the performance very good?
Closed 6 years ago1
Creating norm_info_mgc_lf0_vuv_bap_63_MVN.dat for the Full VCTK dataset
Updated 6 years ago7
How to train the samples obtained from public？too much noise and speed too fast!
Closed 6 years ago1
Pretrained checkpoints for Blizzard 2013 and LJSpeech
Closed 6 years ago1
Full VCTK dataset
Closed 6 years ago1
Blizzard Model
Closed 6 years ago6
From English into Chinese?
Closed 6 years ago1
espeak or espeak-NG?
Closed 6 years ago2
Why scale the outputs by 30 in Decoder.update_buffer?
Updated 6 years ago1
Reproducing the results
Closed 6 years ago9
Error during training
Closed 6 years ago2
Confusion of the Phoneme Generation
Closed 6 years ago1
What is the split ratio between the training set and test set of the VCTK data provided in your project?
Closed 7 years ago2
Training on big dataset.
Closed 7 years ago1
Can loop support parallel training in multiple GPU?
Closed 7 years ago1
' out of memory' accrued when training with single speaker
Closed 7 years ago2
Face 'Not a finite gradient or too big, ignoring.' when training other data.
Closed 7 years ago1
Can we use multi-GPU to make the training faster? How?
Closed 7 years ago2
Error when synthesis wave with the download model.
Closed 7 years ago1
Does anyone successfully train a multi-speaker loop model with his/her own data?
Closed 7 years ago
Training Epochs Number
Closed 7 years ago1