Scheduled generation

Scheduled generation lines up several sets of settings (different models, LoRAs and prompts) and generates them in order automatically. It suits long unattended runs, for example testing a new LoRA with a fixed set of prompts.

Adding to the queue

• Add the current settings: after setting up the generate panel, press “Queue” next to the Generate button. The current model, LoRAs, prompt, settings and count are recorded as one item. Changing the generate panel afterwards doesn’t affect items already queued.
• Add several prompts at once: in “Add several prompts at once” below the queue list, enter one prompt per line and press “Add each line”. Each line becomes an item; everything else uses the generate panel’s current settings.

The queue list

Click “Queue” at the top of the generate panel to open the list.

Button	Description
Start	Generates the items in order
Pause	Pauses after the current image; press “Resume” to continue
Move up, Move down	Change the order
Remove from queue	Deletes the item
Retry failed	Regenerates only the failed images; finished ones aren’t redone
Clear finished	Removes finished items from the list

The list and its progress are kept when you close CoCo. Press “Start” next time to continue where it stopped.

GPU rest

For long runs, choose an option under “GPU rest” in the list:

Option	Description
Off	Generates continuously
Every N images	Pauses for the given number of seconds after every N images
By temperature	When the GPU is above the given temperature, waits for it to cool before the next image

While resting or cooling down, the remaining time or the current temperature is shown above the list.