InvokeAI/ldm/invoke
Lincoln Stein c4fb8e304b fix noisy images at high step counts
At step counts greater than ~75, the ksamplers start producing noisy
images when using the Karras noise schedule. This PR reverts to using
the model's own noise schedule, which eliminates the problem at the
cost of slowing convergence at lower step counts.

This PR also introduces a new CLI `--save_intermediates <n>' argument,
which will save every nth intermediate image into a subdirectory
named `intermediates/<image_prefix>'.

Addresses issue #1083.
2022-10-14 16:19:45 -04:00
..
generator rename all modules from ldm.dream to ldm.invoke 2022-10-08 11:37:23 -04:00
restoration rename all modules from ldm.dream to ldm.invoke 2022-10-08 11:37:23 -04:00
args.py fix noisy images at high step counts 2022-10-14 16:19:45 -04:00
conditioning.py rename all modules from ldm.dream to ldm.invoke 2022-10-08 11:37:23 -04:00
devices.py rename all modules from ldm.dream to ldm.invoke 2022-10-08 11:37:23 -04:00
image_util.py rename all modules from ldm.dream to ldm.invoke 2022-10-08 11:37:23 -04:00
log.py rename all modules from ldm.dream to ldm.invoke 2022-10-08 11:37:23 -04:00
pngwriter.py Add back old dream.py as legacy_api.py 2022-10-12 20:35:56 -04:00
readline.py fix noisy images at high step counts 2022-10-14 16:19:45 -04:00
server_legacy.py Add back old dream.py as legacy_api.py 2022-10-12 20:35:56 -04:00
server.py rename all modules from ldm.dream to ldm.invoke 2022-10-08 11:37:23 -04:00