[Feature Request] save prompt .txt next to previews generated by hypernetwork training

[space-nuko]

Is your feature request related to a problem? Please describe. Ive sometimes got some really good outputs that I want to reproduce from hypernetwork training, but there isn't a .txt file generated for each preview so it's hard to remember what was passed to txt2img to produce them

Describe the solution you'd like Just save the .txt as with normal txt2img for every hypernetwork preview image

Describe alternatives you've considered Saving the hypernetwork preview prompt somewhere by hand

Additional context None

[SeverianVoid]

As commented on the commit this has resulted in the training images generated during training to now respect the "Save images to a sub directory" toggle option so any prompt template that produces different prompts for sampling now floods the sample images folder with sub directories. Probably needs a separate option to enable/disable save to sub directory for training since sub directories are nice for image organization but terrible for comparing images generated during training.

[aleksusklim]

Can we make the same for embedding training too? https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/3186

[timntorres]

As commented on the commit this has resulted in the training images generated during training to now respect the "Save images to a sub directory" toggle option so any prompt template that produces different prompts for sampling now floods the sample images folder with sub directories. Probably needs a separate option to enable/disable save to sub directory for training since sub directories are nice for image organization but terrible for comparing images generated during training.

For posterity: this bug has been addressed by #3625.