Closed eschmidbauer closed 1 month ago
How deep is the E2TTS network in your configuration? There might be difference in loss between depth=4 and depth=16, for example
I'm using the default example shown in train_example.py
- which appears to be depth = 4
Should I stop training and try 16?
This is the configuration in the paper:
was just shared some successful results, so i think this issue can be closed
thanks! i started retraining with latest code this morning
@eschmidbauer The new projection layer should help you size the transformer more appropriately, but report back with results if you can!
I did a few epochs of 100+ hours of LibriTTS-R on a single H100 with a ~100M param model and the loss converged much lower than you reported — it's dataset / training recipe dependent, but I saw ~0.5 loss before I stopped it and visually the generated mel spectrograms looked reasonably accurate.
@eschmidbauer The new projection layer should help you size the transformer more appropriately, but report back with results if you can!
I did a few epochs of 100+ hours of LibriTTS-R on a single H100 with a ~100M param model and the loss converged much lower than you reported — it's dataset / training recipe dependent, but I saw ~0.5 loss before I stopped it and visually the generated mel spectrograms looked reasonably accurate.
Could you share number of heads and layers you used?
this is what im using to train
duration_predictor = DurationPredictor(
transformer = dict(
dim = 512,
depth = 6,
)
)
e2tts = E2TTS(
duration_predictor = duration_predictor,
transformer = dict(
dim = 512,
depth = 12,
skip_connect_type = 'concat'
),
)
optimizer = Adam(e2tts.parameters(), lr=7.5e-5)
trainer = E2Trainer(
e2tts,
optimizer,
num_warmup_steps=20000,
checkpoint_path = 'e2tts.pt',
log_file = 'e2tts.txt'
)
not getting anything useful out of the checkpoint yet but the loss is definitely converging much better
i've started to run the training with total epochs set to 10000. The loss seems to not decrease below 2 after around 40k steps. See the loss graph attached.