-
https://git.ecker.tech/ aims to provide a place to share my efforts while maintaining true ownership of my code, as I do not trust GitHub.
XMR: 4B9TQdkAkBFYrbj5ztvTx89e5LpucPeTSPzemCihdDi9EBnx7btn8RDNZTBz2zihWsjMnDkzn5As1LU6gLv3KQy8BLsZ8SG
- Joined on
2022-10-10
Right, the final loss value's too high, you'll need to train it a lot more, although I guess it's a matter of how nice the dataset is with how fast it'll finetune too, even with similar metrics.
So the training has currently taken up about 400gb of space on the drive (luckily I have the space), but I'm wondering if this is how much will need to be stored long-term for this training to…
It was named as "Do Not Load TTS On Startup", I almost forgot.
Whenever I get a chance I'll have it download the required dvae.pth file when training if it's missing, rather than rely on TTS…
I suppose it has been overtrained and fried; the finetune I just finished still sounds terrible. I'm going to try and generate against older snapshots to see if I can find something sounding decent.
Will take a look when I get a chance.
I had to re-run the finetune on my paperspace instance since I woke up to it halted a small ways through. It looks like this right now so it seems…
Does that seem like an appropriate setup in your opinion?
mmm, yeah, with the default LR rate of 1e-5, it should be at the same point (twice the size the size at half the requested epoch…
Apologies for taking a while to get around to it, but the foundation for has been implemented in commit 7798767fc6a00601b1b2ca487738e9c54a1f4db2. I had a eureka moment earlier about how I can…
Yeah, there was an issue (#73, #74) with haphazardly adding in whisperx as a backend, where it'll break a package DLAS requires. The setup and update-force scripts have been fixed to have DLAS…
Perfect, I was thinking of this earlier with the added complexity and bloated dependencies.