but tbh, same as list.txt val and train and dataset as zip file

Im not entirely sure if i understand you either lol, you mean I should link the model to the colab as zip file?

mmm, i think from what i understand you want continue training or re-train the model?

If anything, I just wanted to ask if anyone else got the same error I had in that other colab i was trying to use https://drive.google.com/file/d/1S4Q85XHhx889hXRpC6U-BKhklQCvu9xe/view Its supposed to train the wavglow vocoder model for tacotron2 seperately, but when i try to run the 2nd block in colab, it ends up spitting an error message relating to installing the necessary components for training lol

well... that's the problem....

i recommend starting with simplified training

it's same training as waveglow/tacotron2

I already did train my file dont worry lol

It gave me a decent model

oh wait you did train model?

I just wanted to improve it with training the waveglow vocoder

Yeah

and from what i understand, you need continue train of it?

No, i want to train the waveglow vocoder model to improve its quality, seperately from the tacotron2 model. Thats why i was looking at that video I linked and I tried to use the Waveglow colab i found

But I cant use it because at the 2nd block in the colab it gives me a error message relating to installing somwthing, and im just asking if someone else gets the same error lol

we prolly don't use that in server tho

so idk how can i help with this

i only know is that pulling dataset (wavs and list.txt) and then train it, then synthesis with finished model (as in drive might go no .zip)

which that synthesis is waveglow

Im just asking if someone gets the same error as I do, all that would take is checking if the colab actually installs the components in the 2nd block of code, but its okay

I'll keep asking

I used a seperate vocoder colab, HiFi-Gan to improve the Vocoder model for my Tacotron2 model, so Im sure the tacotron2 model i trained on uberducks colab is not corrupted or anything but the waveglow colab i tried to use

hence im asking for anyone else who has actually treid using that same waveglow colab

or if someone knows another public waveglow colab link

why the mel spectogram is o 0% for me?

just look

and: Checking Training Files wavs/Born.npy|I am born. [WARNING] does not exist. --------------------------------------------------------------------------- FileNotFoundError Traceback (most recent call last) in () 4 get_ipython().system("sed -i -- 's,.wav|,.npy|,g' {hparams.training_files}; sed -i -- 's,.wav|,.npy|,g' {hparams.validation_files}") 5 ----> 6 check_dataset(hparams) 2 frames /usr/local/lib/python3.7/dist-packages/numpy/lib/npyio.py in load(file, mmap_mode, allow_pickle, fix_imports, encoding) 414 own_fid = False 415 else: --> 416 fid = stack.enter_context(open(os_fspath(file), "rb")) 417 own_fid = True 418 FileNotFoundError: [Errno 2] No such file or directory: 'wavs/Born.npy'

Did you upload your WAVS in the correct path?