MCPcopy Create free account
hub / github.com/MotrixLab/FineMoGen / load_pretrained

Method load_pretrained

mogen/models/rnns/t2m_bigru.py:120–122  ·  view source on GitHub ↗
(self, ckpt_path)

Source from the content-addressed store, hash-verified

118 self.max_text_len = max_text_len
119
120 def load_pretrained(self, ckpt_path):
121 checkpoint = torch.load(ckpt_path, map_location='cpu')
122 self.text_encoder.load_state_dict(checkpoint['text_encoder'])
123
124 def forward(self, text, token, device):
125 B = len(text)

Callers 1

load_pretrainedMethod · 0.45

Calls

no outgoing calls

Tested by

no test coverage detected