Update ddp_train.py
Browse files- ddp_train.py +2 -2
ddp_train.py
CHANGED
|
@@ -190,14 +190,14 @@ def main(config_path):
|
|
| 190 |
for batch in tqdm(val_dataloader, desc=f"Epoch {epoch}/{epochs} [Val]"):
|
| 191 |
batch = [b.to(device) for b in batch]
|
| 192 |
|
| 193 |
-
text_input, text_input_length, mel_input, mel_input_length
|
| 194 |
|
| 195 |
# Forward pass
|
| 196 |
attn_soft, attn_logprob = aligner(spec=mel_input,
|
| 197 |
spec_len=mel_input_length,
|
| 198 |
text=text_input,
|
| 199 |
text_len=text_input_length,
|
| 200 |
-
attn_prior=
|
| 201 |
|
| 202 |
# Calculate loss
|
| 203 |
val_loss = forward_sum_loss(attn_logprob=attn_logprob,
|
|
|
|
| 190 |
for batch in tqdm(val_dataloader, desc=f"Epoch {epoch}/{epochs} [Val]"):
|
| 191 |
batch = [b.to(device) for b in batch]
|
| 192 |
|
| 193 |
+
text_input, text_input_length, mel_input, mel_input_length = batch
|
| 194 |
|
| 195 |
# Forward pass
|
| 196 |
attn_soft, attn_logprob = aligner(spec=mel_input,
|
| 197 |
spec_len=mel_input_length,
|
| 198 |
text=text_input,
|
| 199 |
text_len=text_input_length,
|
| 200 |
+
attn_prior=None)
|
| 201 |
|
| 202 |
# Calculate loss
|
| 203 |
val_loss = forward_sum_loss(attn_logprob=attn_logprob,
|