-
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
c12ada600b
added reset generation settings to default button, revamped utilities tab to double as plain jane voice importer (and runs through voicefixer despite it not really doing anything if your voice samples are already of decent quality anyways), ditched load_wav_to_torch or whatever it was called because it literally exists as torchaudio.load, sample voice is now a combined waveform of all your samples and will always return even if using a latents file
b4ca260de9
added flag to enable/disable voicefixer using CUDA because I'll OOM on my 2060, changed from naively subdividing eavenly (2,4,8,16 pieces) to just incrementing by 1 (1,2,3,4) when trying to subdivide within constraints of the max chunk size for computing voice latents
b16eb99538
history tab doesn't naively reuse the voice dir instead for results, experimental "divide total sound size until it fits under requests max chunk size" doesn't have a +1 to mess things up (need to re-evaluate how I want to calculate sizes of bests fits eventually)
I restored the random voice option, as that got forgotten from adding in a web UI in commit 37d25573accf2dce213cc5ec72c05c4afa02f2b5. It would've been nice when I was testing the colab notebook…
I tried implementing it a while back (as evident with the commented out samplers here, and adding k-diffusion as a…
Right, I forgot a simple cuda=get_device_name() == "cuda".
Added in commit 5e843fe29d886afc8371fa1c343fe70f7bf02cc3.
2427c98333
Implemented kv_cache "fix" (from 1f3c1b5f4a); guess I should find out why it's crashing DirectML backend
b383222be2
Merge pull request 'Download from Gradio' (#31) from Armored1065/tortoise-tts:main into main
Should be fixed in commit a84aaa4f96ac131968b41afda5508e6636f9a03b.
I suppose that would be why some settings seemed to corrupt, and why Listen got set to False for me earlier but I didn't…
409dec98d5
fixed voicefixing not working as intended, load TTS before Gradio in the webui due to how long it takes to initialize tortoise (instead of just having a block to preload it)
b85c9921d7
added button to recalculate voice latents, added experimental switch for computing voice latents
Should be fixed in 2210b49cb6df4a2a919f1773d20ab309d486cb0e. I suppose the issue was that I never actually re-tested computing the latents outside of the CPU, as the DirectML forces it to be done…