Releases: neph1/finetrainers-ui
v0.14.0 Control lora settings
Add settings for control lora training
Add settings for custom models.
I've verified that training runs. But I've not been able to verify results, yet.
v0.12.1 Supports finetrainers v0.1.0
Tested with finetrainers v0.1.0
Removed all validation from the ui, for now.
v0.11.2 - optional delayed start
If you're energy conscious (as in varying prices), or have the machine doing other stuff and want to run training later, you can now enter a delay in minutes to when the training should start.
Tested with finetrainers v0.0.1 (not v0.1.0)
v0.11.1 remove reference to datasettab
Merge pull request #24 from neph1/neph1-patch-1 Update app.py
v0.11.0 add wan to model options
Although the ui supports any model through entering it as a string, this release adds wan as well as cogvideox to the preconfigured options.
Please note that I haven't tested wan.
I also removed diffusion options, since they have changed in latter versions of finetrainers (Issue #20 ) . Will re add later when I have time
Also removed the lora tab since it was outdated. Better options exist for turning the lora comfy compatible.
Tested with finetrainers v0.0.1 / a-r-r-o-w/finetrainers@41dd338
Edit: Initial commit tag was wrong
v0.10.1 add missing settings + hunyuan_video
Tested with finetrainers commit: 3eb9f8a6c4541e58ec5b3c7522ab003d2aa2d5f9
What's Changed
Full Changelog: v0.10.0...v0.10.1
v0.10.0 fp8 training
Tested with finetrainers commit: 4271b4e436b6057c34c8cc155b997bb63f65fb00
What's Changed
Full Changelog: v0.9.3...v0.10.0