14 lines
364 B
Python
14 lines
364 B
Python
class SpeechEncoder(object):
|
|
def __init__(self, vec_path="pretrain/checkpoint_best_legacy_500.pt", device=None):
|
|
self.model = None # This is Model
|
|
self.hidden_dim = 768
|
|
pass
|
|
|
|
|
|
def encoder(self, wav):
|
|
"""
|
|
input: wav:[signal_length]
|
|
output: embedding:[batchsize,hidden_dim,wav_frame]
|
|
"""
|
|
pass
|