MCPcopy Create free account
hub / github.com/MoonInTheRiver/DiffSinger / build_model

Method build_model

tasks/tts/fs2.py:68–73  ·  view source on GitHub ↗
(self)

Source from the content-addressed store, hash-verified

66 self.model = FastSpeech2(self.phone_encoder)
67
68 def build_model(self):
69 self.build_tts_model()
70 if hparams['load_ckpt'] != '':
71 self.load_ckpt(hparams['load_ckpt'], strict=True)
72 utils.print_arch(self.model)
73 return self.model
74
75 def _training_step(self, sample, batch_idx, _):
76 loss_output = self.run_model(self.model, sample)

Callers

nothing calls this directly

Calls 2

build_tts_modelMethod · 0.95
load_ckptMethod · 0.80

Tested by

no test coverage detected