Comments (20)
For those still struggling with the initial problem:
The "case" the script asks for is your name for the case, passed as an argument. e.g.
python train1.py test
where "test" can be anything. I suggest naming it after the dataset you use, such as
python train1.py TIMIT
if you're using the TIMIT dataset.
from deep-voice-conversion.
ok...I solve that problem now... but now I find the training stop at 0b [00:00 , ?b/2]
from deep-voice-conversion.
You don't have Timit datasets?
from deep-voice-conversion.
You can get the trained weights here so you don't have to worry about Train1.
from deep-voice-conversion.
mfcc, ppg = get_batch(model.mode, model.batch_size)
File "G:\code\python\myfile\ASR\deep-voice-conversion-master\data_load.py", line 205, in get_batch
mfcc, ppg = list(map(get_zero_padded, list(zip([get_mfccs_and_phones(w, hp_default.sr) for w in target_wavs]))))
File "G:\code\python\myfile\ASR\deep-voice-conversion-master\data_load.py", line 205, in
mfcc, ppg = list(map(_get_zero_padded, list(zip([get_mfccs_and_phones(w, hp_default.sr) for w in target_wavs]))))
File "G:\code\python\myfile\ASR\deep-voice-conversion-master\data_load.py", line 25, in get_mfccs_and_phones
wav, sr = librosa.load(wav_file, sr=sr)
File "G:\anaconda\lib\site-packages\librosa\core\audio.py", line 112, in load
exec(compile(f.read(), filename, 'exec'), namespace)
File "G:/code/python/myfile/ASR/deep-voice-conversion-master/train1.py", line 117, in
train(logdir=logdir, hparams = hp)
File "G:/code/python/myfile/ASR/deep-voice-conversion-master/train1.py", line 65, in train
with audioread.audio_open(os.path.realpath(path)) as input_file:
File "G:\anaconda\lib\site-packages\audioread_init.py", line 114, in audio_open
raise NoBackendError()
audioread.NoBackendError
execfile(filename, namespace)
File "G:\anaconda\lib\site-packages\spyder\utils\site\sitecustomize.py", line 102, in execfile
Traceback (most recent call last):
File "", line 1, in
File "G:\anaconda\lib\site-packages\spyder\utils\site\sitecustomize.py", line 880, in runfile
from deep-voice-conversion.
raise NoBackendError()
audioread.NoBackendError
It raise NobackendError...... @JiaYK @VictoriaBentell Do you know anything about it? Any comments would be much appreciated..... Thanks for your answer... I haven't put my Timit dataset in it last time
from deep-voice-conversion.
According to this thread you might want to try installing ffmpeg.
from deep-voice-conversion.
It's better to work under Linux. Tensorflow-gpu does not support Python2.7.
from deep-voice-conversion.
@JiaYK @VictoriaBentell Dear friends, it this means I success training now? the epoch and step goes strange
INFO:tensorflow:Restoring parameters from ./logdir/default1/train1\epoch_24_step_24
Model loaded. mode: train1, model_name: epoch_24_step_24
acc: 0.0425952
loss: 3.95336
WARNING:tensorflow:From G:\code\python\myfile\ASR\deep-voice-conversion-master\models.py:73: arg_max (from tensorflow.python.ops.gen_math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use argmax
instead
INFO:tensorflow:Restoring parameters from ./logdir/default1/train1\epoch_24_step_24
Model loaded. mode: train1, model_name: epoch_24_step_24
acc: 0.0476528
loss: 3.8893
WARNING:tensorflow:From G:\code\python\myfile\ASR\deep-voice-conversion-master\models.py:73: arg_max (from tensorflow.python.ops.gen_math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use argmax
instead
INFO:tensorflow:Restoring parameters from ./logdir/default1/train1\epoch_26_step_26
Model loaded. mode: train1, model_name: epoch_26_step_26
acc: 0.0535262
loss: 3.91394
WARNING:tensorflow:From G:\code\python\myfile\ASR\deep-voice-conversion-master\models.py:73: arg_max (from tensorflow.python.ops.gen_math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use argmax
instead
INFO:tensorflow:Restoring parameters from ./logdir/default1/train1\epoch_26_step_26
Model loaded. mode: train1, model_name: epoch_26_step_26
acc: 0.0459833
loss: 3.98388
from deep-voice-conversion.
Those warnings are normal, so if you just let it keep training you should be fine.
from deep-voice-conversion.
How did you solve the "argument" problem in the first case?
I am facing the same issue.
from deep-voice-conversion.
@VictoriaBentell did y manage to get the whole pipeline working (with convert) ?
Is 12 epochs of train1 enough ?
from deep-voice-conversion.
I'm surprised this isn't in the docs. It was confusing. Thanks to @CIDFarwin and #35 (comment)
from deep-voice-conversion.
test what erver
Traceback (most recent call last):
File "train1.py", line 76, in
print('case: {}, logdir: {}'.format(args.case1, args.case, logdir_train1))
from deep-voice-conversion.
You can get the trained weights here so you don't have to worry about Train1.
thank you
from deep-voice-conversion.
reading hparams.py and u can find your answer
default_hp = load_hparam(default_file) user_hp = load_hparam(user_file) hp_dict = Dotdict(merge_dict(user_hp[case], default_hp) if case in user_hp else default_hp)
here the author merged two dicts: user_hp[case] & default_hp , when u dig into those .yaml files in hparams dir &merge_dict() function in hparams.py ,u can find that :case may refer to different attempt on training hyperparamters(case just like a Fine-tuned version number) ,as u can see the merge_dict() fn replace params in default_hp with same named params in user_hp[case]
from deep-voice-conversion.
raise NoBackendError()
audioread.NoBackendError
It raise NobackendError...... @JiaYK @VictoriaBentell Do you know anything about it? Any comments would be much appreciated..... Thanks for your answer... I haven't put my Timit dataset in it last time
i meet the same problem ,may i ask u for the answer?
from deep-voice-conversion.
my code with train1.py stunk here, anyone konw how to solve it?
from deep-voice-conversion.
You can get the trained weights here so you don't have to worry about Train1.
hello,do you still have the trained weights from drive.google.com?I can't open the url and also can't download the weights.Could you send the weights to me or download then send to me?thanks.@VictoriaBentell
from deep-voice-conversion.
You can get the trained weights here so you don't have to worry about Train1.
hello,do you still have the trained weights from drive.google.com?I can't open the url and also can't download the weights.Could you send the weights to me or download then send to me?thanks.@VictoriaBentell
It works for me in an incognito window on Windows 10 in Google Chrome. It's been a while but I think the file you should be looking at is "checkpoint"
from deep-voice-conversion.
Related Issues (20)
- python3.6 Successful Attempt HOT 2
- InvalidArgumentError while running tran1.py HOT 2
- other dataset try,eg ljspeech
- Does anyone succesfully used one gpu with docker? HOT 1
- TIMIT DATASET HOT 1
- Voice quality not good HOT 8
- prebuilt windows release? HOT 1
- Problems I encountered when running train1 HOT 2
- ------Model Request For Train2-----
- why my tensorboard audio tab is giving only 2 sec of output???
- Why does acc fluctuate when I train with my own corpus? HOT 1
- structure of models?
- Dataset link not working
- where can I find the colab version of this repo? HOT 1
- LRU_cache error
- which part is that compute the MCEPs
- The successful replication of the PyTorch version of this project
- logdir_train1 = '{}/{}/train1'.format(hp.logdir_path, args.case) KeyError: 'logdir_path'
- stuck on train1 without error
- Train1 code is failed with AttributeError: 'Net1' object has no attribute 'inputs'
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 deep-voice-conversion.