building something like that was what originally motivated uberduck

Oh sweet, I will let you know how it goes! 😋

Ok got it downloaded finally but couldn't get the container to start 😅

state_dict = torch.load(r"C:\Users\superuser\Desktop\DeepLearningExamples-master\PyTorch\SpeechSynthesis\Tacotron2\nvidia_tacotron2pyt_fp16.pt", map_location=torch.device('cpu'))['state_dict']

is that right?

yep, looks right

But error come

An massive error

what's the error?

its super massive error

@zwf I fixed my website and how can I censor some words?

why did you write code in the brainfuck language

oh its an error

Who?

it was a joke, i realized brainfuck was a coding language

@User however I did notice that that's a very short utterance by encoder length

yes

i will try batch size 2

what batch size you use?

like 30

can we use batch size 60 for 48gb gpu memory

?

yes, but I don't think you have 48GB GPU memory? where are you seeing that number?

i have 8gb gpu memory

but its amd

can we use training for 6gb vram?

but how much batch size does it need?

the way to set batch size is to just try with a higher number, and decrease it until you find something that doesn't run out of memory

are you still transcribing the Phil Swift?

oh wait