search

tugstugi / dl-colab-notebooks

Try out deep learning models online on Google Colab
1.72k stars 465 forks source link

Add Real-Time-Voice-Cloning? #18

Closed htadashi closed 4 years ago

htadashi commented 4 years ago

Thanks for this fantastic resource. I think this collab on real time voice cloning could also be a nice addition to the list: https://colab.research.google.com/drive/1ts-KRxOuuAHmb-AtjU1XTM_lIFOepAN2

tugstugi commented 4 years ago

tnx, I have updated the demo notebook little bit and uploaded it: https://colab.research.google.com/github/tugstugi/dl-colab-notebooks/blob/master/notebooks/RealTimeVoiceCloning.ipynb

400lbhacker commented 3 years ago

last step, error: "operands could not be broadcast together with shapes (2200,) (4000,) (2200,)" is this due to numpy version or python version(2/3) thanks

tugstugi commented 3 years ago

@400lbhacker fixed.

aakashkumar6868 commented 3 years ago

Hey can u do it for adding paragraphs.

400lbhacker commented 3 years ago

Hey tugstugi I just wanted to say thankyou for making this wonderful project, and for also resolving that issue. That is the fastest fix responce time ive ever had on github. keep up the wonderful project

400lbhacker commented 3 years ago

I really want to talk to you im impressed with alot of your projects im new to github still learning how to use it, my email is here: josepherickson135@gmail.com my facebook is: https://www.facebook.com/profile.php?id=100015734717430

himanshuragtah1 commented 2 years ago

Hey can u do it for adding paragraphs.

@400lbhacker fixed.

@tugstugi error: "operands could not be broadcast together with shapes (2200,) (4000,) (2200,)" occurring again. How can this be fixed?

deanrussell commented 2 years ago

Hi @tugstugi. On colab, I'm getting two errors in the final "Synthesize" step:

FileNotFoundError Traceback (most recent call last) [<ipython-input-4-6e155181f8d0>](https://localhost:8080/#) in <module>()

and

FileNotFoundError: [Errno 2] No such file or directory: 'Real-Time-Voice-Cloning/saved_models/default/synthesizer.pt'

Any thoughts on what I'm doing wrong?

tugstugi commented 2 years ago

@deanrussell Currently download from Google Drive doesn't work anymore, as long gdown doesn't fix it, we can't do anything, see: https://github.com/wkentaro/gdown/issues/162

tugstugi commented 2 years ago

@deanrussell you can try again, it is now hopefully fixed.

deanrussell commented 2 years ago

@deanrussell you can try again, it is now hopefully fixed.

Thank you, @tugstugi! It works.

Cybernite21 commented 2 years ago

Hi, this is a really great tool, but if I were to disconnect and reconnect, is there a way to reload the state the ai was in? So i do not have to re upload audio and retrain?

KaliHUB10 commented 2 years ago

THANK YOU

KaliHUB10 commented 2 years ago

Thank YOU

12 Ağu 2022 Cum 14:51 tarihinde rwconsultancy @.***> şunu yazdı:

I must be doing something wrong but I don't know what. Although my own recording is loud and clear, the generated voice is muffled and garbled - it rather sounds as if I'm talking gibberish with a sock in my mouth.

The installation moaned about needing a later version of numpy, so I gave it that, but that didn't change anything. I also had to change the line '%tensorflow_version 1.x' to '%tensorflow_version 2.x', otherwise it wouldn't install; I don't know if that has made a difference. There were no other errors.

Be gentle with me - I really have very little clue what I'm doing! I'm just looking for an easy way to clone my own voice. Thanks!

— Reply to this email directly, view it on GitHub https://github.com/tugstugi/dl-colab-notebooks/issues/18#issuecomment-1213030872, or unsubscribe https://github.com/notifications/unsubscribe-auth/AWXXAC77WBBK3PXFQTZDQOTVYY3FNANCNFSM4KJUQMYQ . You are receiving this because you commented.Message ID: @.***>

souvikg544 commented 2 years ago

Will giving big audio files of maybe 5 hours help in getting a good output?

lucidigital commented 2 years ago

tnx, I have updated the demo notebook little bit and uploaded it: https://colab.research.google.com/github/tugstugi/dl-colab-notebooks/blob/master/notebooks/RealTimeVoiceCloning.ipynb

Google Colab said they don't support TensorFlow 1.x anymore, and require to update TensorFlow 2. Colab doc becomes useless :( Can you update it? Thanks

jujpenabe commented 1 year ago

Must be upgraded to tensorflow 2 for compatibility with colab.

AlbinoGeek commented 1 year ago

For all the people mentioning tensorflow 2 support....

Just delete the line that mentions tensorflow 1.x -- it works.

@jujpenabe @lucidigital

Abbienew commented 1 year ago

My first attempt was successful, but the subsequent executions were unsuccessful and produced errors ModuleNotFoundError: No module named 'unidecode' ; NameError: name 'encoder' is not defined;" NameError: name 'encoder' is not defined "

any solutions? tips?

imvbhuvan commented 1 year ago

TypeError Traceback (most recent call last) in <cell line: 23>() 28 #button = widgets.Button(description="Upload Voice File") 29 #button.on_click(_upload_audio) ---> 30 _upload_audio("")

2 frames /content/Real-Time-Voice-Cloning/encoder/audio.py in preprocess_wav(fpath_or_wav, source_sr, normalize, trim_silence) 40 # Resample the wav if needed 41 if source_sr is not None and source_sr != sampling_rate: ---> 42 wav = librosa.resample(wav, source_sr, sampling_rate) 43 44 # Apply the preprocessing: normalize volume and shorten long silences

TypeError: resample() takes 1 positional argument but 3 were given

I am this error constantly any tips on how to resolve this @tugstugi

roy-sub commented 6 months ago

Hey imvbhuvan where you able to fix the TypeError: resample() takes 1 positional argument but 3 were given.

If yes kindly provide the solution !