justinjohn0306/fakeyou-tacotron2-notebook

Tacotron2 Training Notebook for FakeYou.com

FakeYou-Tacotron2-Notebooks

Tacotron2 Training and Synthesis Notebooks for FakeYou.com

Google Colab

Training Notebook (ENG): and follow the instructions.

Synthesis Notebook (CPU): and follow the instructions.

Synthesis Notebook (GPU): and follow the instructions.

Spanish Training and Synthesis nbs

Training Notebook (ES): and follow the instructions.

Synthesis Notebook (CPU): and follow the instructions.

Bonus

Vocoder Fine-tuning Notebook: and follow the instructions.

fakeyou-tacotron2-notebook's Issues

Checking Training Files [WARNING] does not exist.

I tried every step 3x now but I am still getting the same error. Am I doing something wrong?

Just have a small request

Can you please give me a link to the jarvis voice model... i am running into errors while training... would be really helpful since im making my own Ai Assistant. Thank you

num_samples should be a positive integer value, but got num_samples=0

I'm getting an error at the begin of training: num_samples should be a positive integer value, but got num_samples=0.
How to fix it?

AssertionError when trying to generate GTA spectrograms

Hello.
I am a newbie on faking voices and I am trying to fine-tune HiFi-GAN for a personal project using this colab notebook
When trying to generate GTA spectrograms:

"!python GTA.py -c tacomodel -o GTAOutput" an AssertionError appears

I have tried with a large dataset and a small one with 30 phrases.
I have checked puntuation, samplerate and mono sound for wav files.
Also checked train_filelist.txt, val_filelist.txt using a small dataset to avoid errors and still gettting "AssertionError"

Do you think it is a dataset error or something inside the code?

About Jarvis Voice

I need a little bit of your help. I need JARVIS voice for my project. I am here to ask how can I integrate his voice in it so that it runs perfectly even when it is offline

FileNotFoundError: [Errno 2] No such file or directory: '\ufeffwavs/01.npy'

Please delete this issue, I resolved it.

Link down, reupload spanish base models pls

can you reupload Spanish-male or Spanish-female base models? i can't find them anywhere, pls

Using my own trained model as pre-trained model

how can i using my own trained model as pre-trained model in this collab notebook? which part should I change?

ValueError: Cannot load file containing pickled data when allow_pickle=False

Getting this error when trying to train the model. Runtime error when first epoch starts.
Changed the
melspec = torch.from_numpy(np.load(filename))
to
melspec = torch.from_numpy(np.load(filename, allow_pickle=True))
which still failed.

Can't convert WAVs to Mel-Spectrograms

I can't get this working, any suggestions?

Cannot detect wav files for conversion to mel spectrogram

Hi! I am getting this error where the create_mel() function is unable to detect the .wav file for some reason.

This is despite being able to extract the .wav files from zip.

In addition, here are my audacity settings for conversion of the audio files: 22050 Hz video rate and mono setting

Do you happen to know what might be the problem?

not downloading wavs folder

I ran the training notebook but it's not downloading the wavs folder. WHAT IS GOING ON?! WHY IS IT SKIPPING A STEP THAT'S SO EASY TO DO!??!?!?!?!?!?!??! >8(

Encoder Timesteps only to 40???

I was training a model with 152 voice lines and I got like a 100 Encoder Timesteps on the graphs while training but I am doing another model with 182 voice lines and I get only 40. Any way to fix???