I have trained tacotron2 model with recipes provided, how to use it?

[bowiesimon9]

I write a simple code to test it , but get error.
my code:
I just copy the code from the example:
https://colab.research.google.com/drive/1mhvu5eyEzNBBhNC6sknG-kgD2ZQ6NEyI?usp=sharing#scrollTo=0W6oPmoajJl-

import torchaudio
import speechbrain
from speechbrain.pretrained import Tacotron2

tacotron2 = Tacotron2.from_hparams(source="speechbrain/tts-tacotron2-ljspeech", savedir=".")

mel_output, mel_length, alignment = tacotron2.encode_text("This is a test.")
print(mel_output.shape)

hyperparams.yaml :

#model
model: !new:speechbrain.lobes.models.Tacotron2.Tacotron2

#optimizer
  mask_padding: true
  n_mel_channels: 80
  #symbols
  n_symbols: 148
  symbols_embedding_dim: 512
  #encoder
  encoder_kernel_size: 5
  encoder_n_convolutions: 3
  encoder_embedding_dim: 512
  #attention
  attention_rnn_dim: 1024
  attention_dim: 128
  #attention location
  attention_location_n_filters: 32
  attention_location_kernel_size: 31
  #decoder
  n_frames_per_step: 1
  decoder_rnn_dim: 1024
  prenet_dim: 256
  max_decoder_steps: 1000
  gate_threshold: 0.5
  p_attention_dropout: 0.1
  p_decoder_dropout: 0.1
  #postnet
  postnet_embedding_dim: 512
  postnet_kernel_size: 5
  postnet_n_convolutions: 5
  decoder_no_early_stopping: false

text_to_sequence: !name:speechbrain.utils.text_to_sequence.text_to_sequence

modules:
    model: !ref <model>

pretainer: !new:speechbrain.utils.parameter_transfer.Pretainer
    loadables:
        model: !ref <model>

I have put the model.ckpt and hyperparams.yaml to current folder.

and the error:

python3 sp_local_m01.py
Traceback (most recent call last):
  File "/home/bowie/myTTS/code/speechbrain/sp_local_m01.py", line 5, in <module>
    tacotron2 = Tacotron2.from_hparams(source="speechbrain/tts-tacotron2-ljspeech", savedir=".")
  File "/home/bowie/.local/lib/python3.10/site-packages/speechbrain/pretrained/interfaces.py", line 424, in from_hparams
    hparams = load_hyperpyyaml(fin, overrides)
  File "/home/bowie/.local/lib/python3.10/site-packages/hyperpyyaml/core.py", line 188, in load_hyperpyyaml
    hparams = yaml.load(yaml_stream, Loader=loader)
  File "/home/bowie/.local/lib/python3.10/site-packages/yaml/__init__.py", line 81, in load
    return loader.get_single_data()
  File "/home/bowie/.local/lib/python3.10/site-packages/ruamel/yaml/constructor.py", line 114, in get_single_data
    return self.construct_document(node)
  File "/home/bowie/.local/lib/python3.10/site-packages/ruamel/yaml/constructor.py", line 118, in construct_document
    data = self.construct_object(node)
  File "/home/bowie/.local/lib/python3.10/site-packages/ruamel/yaml/constructor.py", line 145, in construct_object
    data = self.construct_non_recursive_object(node)
  File "/home/bowie/.local/lib/python3.10/site-packages/ruamel/yaml/constructor.py", line 186, in construct_non_recursive_object
    for _dummy in generator:
  File "/home/bowie/.local/lib/python3.10/site-packages/ruamel/yaml/constructor.py", line 629, in construct_yaml_map
    value = self.construct_mapping(node)
  File "/home/bowie/.local/lib/python3.10/site-packages/ruamel/yaml/constructor.py", line 427, in construct_mapping
    return BaseConstructor.construct_mapping(self, node, deep=deep)
  File "/home/bowie/.local/lib/python3.10/site-packages/ruamel/yaml/constructor.py", line 242, in construct_mapping
    value = self.construct_object(value_node, deep=deep)
  File "/home/bowie/.local/lib/python3.10/site-packages/ruamel/yaml/constructor.py", line 145, in construct_object
    data = self.construct_non_recursive_object(node)
  File "/home/bowie/.local/lib/python3.10/site-packages/ruamel/yaml/constructor.py", line 181, in construct_non_recursive_object
    data = constructor(self, tag_suffix, node)
  File "/home/bowie/.local/lib/python3.10/site-packages/hyperpyyaml/core.py", line 472, in _construct_object
    raise ImportError("There is no such class as %s" % callable_string)
ImportError: There is no such class as speechbrain.utils.parameter_transfer.Pretainer

the relative folder:

[BenoitWang]

Hi @bowiesimon9, this seems an environment issue to me. Normally if you follow this pipeline it should be ok. You'll only need to change the huggingface link to your local path, and maybe change your repo name (do not name it */speechbrain), I guess that is why it says There is no such class as speechbrain.utils.parameter_transfer.Pretainer, it looks into your local repo.

[bowiesimon9]

thank you for your reply. it is my fault, I have some spelling error in hyperparams.yaml.