提交 9dad6c3d 编写于 作者: C chenfeiyu

fix synthesis for transformerTTS and FastSpeech, use int64 explicitly

上级 b7c584e2
......@@ -83,8 +83,8 @@ def synthesis(text_input, args):
pos_text = np.arange(1, text.shape[1] + 1)
pos_text = np.expand_dims(pos_text, axis=0)
text = dg.to_variable(text)
pos_text = dg.to_variable(pos_text)
text = dg.to_variable(text).astype(np.int64)
pos_text = dg.to_variable(pos_text).astype(np.int64)
_, mel_output_postnet = model(text, pos_text, alpha=args.alpha)
......
......@@ -92,15 +92,17 @@ def synthesis(text_input, args):
model_vocoder.eval()
# init input
text = np.asarray(text_to_sequence(text_input))
text = fluid.layers.unsqueeze(dg.to_variable(text), [0])
text = fluid.layers.unsqueeze(dg.to_variable(text).astype(np.int64), [0])
mel_input = dg.to_variable(np.zeros([1, 1, 80])).astype(np.float32)
pos_text = np.arange(1, text.shape[1] + 1)
pos_text = fluid.layers.unsqueeze(dg.to_variable(pos_text), [0])
pos_text = fluid.layers.unsqueeze(
dg.to_variable(pos_text).astype(np.int64), [0])
pbar = tqdm(range(args.max_len))
for i in pbar:
pos_mel = np.arange(1, mel_input.shape[1] + 1)
pos_mel = fluid.layers.unsqueeze(dg.to_variable(pos_mel), [0])
pos_mel = fluid.layers.unsqueeze(
dg.to_variable(pos_mel).astype(np.int64), [0])
mel_pred, postnet_pred, attn_probs, stop_preds, attn_enc, attn_dec = model(
text, mel_input, pos_text, pos_mel)
mel_input = fluid.layers.concat(
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册