Comments (13)
@xcmyz Great implementation, Is anyway to train transfomer first so that later I used transformer itself to generate alignement and the then train FastSpeech on it this way we can achieve sequence-level knowledge distillation
as mentioned in paper.
So basically I like to train transformer only first, is any way possible using this code itself for training.
Meanwhile I will write a training script with myself for train transformer just curious to ask is it possible through in this code.
from fastspeech.
Got following error after running python preprocess.py
Traceback (most recent call last):
File "preprocess.py", line 59, in <module>
main()
File "preprocess.py", line 52, in main
_, _, D = load_data(character, mel_gt_target, tacotron2)
File "E:\Dev\FastSpeech\utils.py", line 160, in load_data
D = get_D(alignment)
File "E:\Dev\FastSpeech\utils.py", line 79, in get_D
max_index = alignment[i].tolist().index(alignment[i].max())
ValueError: nan is not in list
Problem with alignment, so I printed it and got :
Align : [[nan nan nan ... nan nan nan]
[nan nan nan ... nan nan nan]
[nan nan nan ... nan nan nan]
...
[nan nan nan ... nan nan nan]
[nan nan nan ... nan nan nan]
[nan nan nan ... nan nan nan]]
Are you aware of this kind of issue?
from fastspeech.
Got following error after running
python preprocess.py
Traceback (most recent call last): File "preprocess.py", line 59, in <module> main() File "preprocess.py", line 52, in main _, _, D = load_data(character, mel_gt_target, tacotron2) File "E:\Dev\FastSpeech\utils.py", line 160, in load_data D = get_D(alignment) File "E:\Dev\FastSpeech\utils.py", line 79, in get_D max_index = alignment[i].tolist().index(alignment[i].max()) ValueError: nan is not in list
Problem with alignment, so I printed it and got :
Align : [[nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan] ... [nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan]]
Are you aware of this kind of issue?
What pretrained model did you use? BTW, you can use alignment.zip
.
from fastspeech.
@xcmyz Great implementation, Is anyway to train transfomer first so that later I used transformer itself to generate alignement and the then train FastSpeech on it this way we can achieve
sequence-level knowledge distillation
as mentioned in paper.
So basically I like to train transformer only first, is any way possible using this code itself for training.
Meanwhile I will write a training script with myself for train transformer just curious to ask is it possible through in this code.
You need to modify utils.py
.
from fastspeech.
@xcmyz thanks will try... If it's possible can you please upload Fastspeech pretrain on google drive
or dropbox
etc because as I am not based in China I am not able to download it from Baidu.
from fastspeech.
@xcmyz thanks will try... If it's possible can you please upload Fastspeech pretrain on
google drive
ordropbox
etc because as I am not based in China I am not able to download it from Baidu.
Sorry.., You know, I am in China and my vpn isn't stable enough to put this model in Google drive...
from fastspeech.
@xcmyz ,hello,when i train at the step of 272885,i cannot have a good result,the wav file has noise,what is wrong with that? How many steps should i train to have a good result,thank you .
from fastspeech.
@xcmyz ,hello,when i train at the step of 272885,i cannot have a good result,the wav file has noise,what is wrong with that? How many steps should i train to have a good result,thank you .
>= 100000
from fastspeech.
I get the same error. The pretrained model download from nvidia. Please tell me how to solve this problem? @xcmyz
from fastspeech.
Use learning rate schedule.
from fastspeech.
运行后出现以下错误
python preprocess.py
Traceback (most recent call last): File "preprocess.py", line 59, in <module> main() File "preprocess.py", line 52, in main _, _, D = load_data(character, mel_gt_target, tacotron2) File "E:\Dev\FastSpeech\utils.py", line 160, in load_data D = get_D(alignment) File "E:\Dev\FastSpeech\utils.py", line 79, in get_D max_index = alignment[i].tolist().index(alignment[i].max()) ValueError: nan is not in list
对齐问题,所以我将其打印并得到:
Align : [[nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan] ... [nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan]]
您知道这种问题吗?
did you solve the problem? I have the same bug
from fastspeech.
运行后出现以下错误
python preprocess.py
Traceback (most recent call last): File "preprocess.py", line 59, in <module> main() File "preprocess.py", line 52, in main _, _, D = load_data(character, mel_gt_target, tacotron2) File "E:\Dev\FastSpeech\utils.py", line 160, in load_data D = get_D(alignment) File "E:\Dev\FastSpeech\utils.py", line 79, in get_D max_index = alignment[i].tolist().index(alignment[i].max()) ValueError: nan is not in list
对齐问题,所以我将其打印并得到:
Align : [[nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan] ... [nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan] [nan nan nan ... nan nan nan]]
您知道这种问题吗?
did you solve the problem? I have the same bug
No, maybe you can use alignment.zip directly or check the code about how to get alignment.
from fastspeech.
getting same error, i cannot use alignment from alignment,zip as i need alignment with only 20 mels.
so that i can integrate with lpcnet
from fastspeech.
Related Issues (20)
- RuntimeError: shape '[1, 1, 155520]' is invalid for input of size 311040 [custom data training] HOT 3
- training stops in few seconds and no checkpoint file created HOT 3
- error in new commit HOT 9
- How to get alignment? HOT 8
- Preprocess.py got stuck: Tried to debug
- Have anyone tried using LSTM to replace FFT block?
- What is the difference between postnet and CBHG? HOT 3
- 请问训练多久得到的pretrain model呢?然后,请问使用了多少GPU呢? HOT 1
- Resume training from checkpoint result in NaN? HOT 1
- How to extract alignment from tacotron2? HOT 6
- long int 转换成float erro
- denoiser HOT 1
- some question about squeezewave denoiser
- onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Load model from fastwave.onnx failed:Type Error: Type parameter (T) bound to different types (tensor(double) and tensor(float) in node ().
- Expected object of backend CUDA but got backend CPU for argument #3 'index' HOT 1
- duration loss calulated in log domain or linear domain
- wav in chinese HOT 1
- Error when training new model for another language
- BUG:OSError: sndfile library not found HOT 1
- RuntimeError: stack expects each tensor to be equal size, but got [40, 240] at entry 0 and [78, 202] at entry 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from fastspeech.