@wonwooo

[CHK-0000]

Can you provide the training code for that model?

---

import torch

from naturalspeech2_pytorch import Trainer, EncodecWrapper, Model, NaturalSpeech2, SpeechPromptEncoder

codec = EncodecWrapper()

def main():
model = Model(
dim = 128,
depth = 6,
dim_prompt = 512,
cond_drop_prob = 0.25,
condition_on_prompt = True
)

diffusion = NaturalSpeech2(
    model = model,
    codec = codec,
    timesteps = 50
)

raw_audio = torch.randn(4, 327680)
prompt = torch.randn(4, 32768)

text = torch.randint(0, 100, (4, 100))
text_lens = torch.tensor([100, 50 , 80, 100])

# forwards and backwards

loss = diffusion(
    audio = raw_audio,
    text = text,
    text_lens = text_lens,
    prompt = prompt,
    )

loss.backward()

# after much training

generated_audio = diffusion.sample(
    length = 1024,
    text = text,
    prompt = prompt,
    )

trainer = Trainer(
    diffusion_model = diffusion,
    folder = 'C:\\naturalspeech2-pytorch\\ansunghun',
    train_batch_size = 16,
    gradient_accumulate_every = 2,
    train_num_steps = 5,
    save_and_sample_every = 100,
)

trainer.train()
trainer.save_checkpoint('C:\\naturalspeech2-pytorch\\ansunghun\\checkpoint.pt')

if name == 'main':
from multiprocessing import freeze_support
freeze_support()
main()

---

An error occurs in that code.

---

Traceback (most recent call last):
File "test.py", line 62, in
main()
File "test.py", line 56, in main
trainer.train()
File "C:\naturalspeech2-pytorch\naturalspeech2_pytorch\naturalspeech2_pytorch.py", line 1875, in train
loss = self.model(data)
File "C:\Users\user.conda\envs\svc\lib\site-packages\torch\nn\modules\module.py", line 1511, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "C:\Users\user.conda\envs\svc\lib\site-packages\torch\nn\modules\module.py", line 1520, in _call_impl
return forward_call(*args, **kwargs)
File "C:\naturalspeech2-pytorch\naturalspeech2_pytorch\naturalspeech2_pytorch.py", line 1522, in forward
text_max_length = text.shape[-1]
AttributeError: 'NoneType' object has no attribute 'shape'