d38e560e42
Implement some rethinking about LoRA system
...
1. Add an option to allow users to use UNet in fp8/gguf but lora in fp16.
2. All FP16 loras do not need patch. Others will only patch again when lora weight change.
3. FP8 unet + fp16 lora are available (somewhat only available) in Forge now. This also solves some “LoRA too subtle” problems.
4. Significantly speed up all gguf models (in Async mode) by using independent thread (CUDA stream) to compute and dequant at the same time, even when low-bit weights are already on GPU.
5. View “online lora” as a module similar to ControlLoRA so that it is moved to GPU together with model when sampling, achieving significant speedup and perfect low VRAM management simultaneously.
2024-08-19 04:31:59 -07:00
9c5038c766
Add --forge-ref-a1111-home cmd arg to reference existing A1111 checkout ( #203 )
...
* Add --forge-ref-a1111-home cmd arg to reference existing A1111 checkout
* nit
2024-02-11 21:22:31 -05:00
d0026da483
add --dump-sysinfo, a cmd arg to dump limited sysinfo file at startup
2023-08-30 19:48:47 +03:00
c76a30af41
more info for startup timings
2023-07-22 13:49:29 +03:00
90eb731ff1
start timer early anyway
2023-07-22 12:21:05 +03:00
33694baea1
avoid importing timer when it is not strictly needed
2023-07-21 17:15:44 +08:00
fc3bdf8c11
Merge branch 'dev' into move-start-timer
2023-07-19 10:33:31 +08:00
136c8859a4
add backwards compatibility --lyco-dir-backcompat option, use that for LyCORIS directory instead of hardcoded value
...
prevent running preload.py for disabled extensions
2023-07-18 20:11:30 +03:00
3c570421d3
move start timer
2023-07-18 19:00:16 +08:00
f9fe5e5f9d
reworking launch.py: add references to renamed file
2023-05-21 16:27:34 +03:00
4b07984d1b
reworking launch.py: rename
2023-05-21 16:27:34 +03:00
31545abe14
add DPM-Solver++(2M) SDE from new k-diffusion
2023-05-21 07:31:51 +03:00
793a491923
Overhaul tests to use py.test
2023-05-19 17:42:34 +03:00
3885f8a63e
Merge pull request #10381 from AUTOMATIC1111/minor-fix
...
Minor fix
2023-05-18 17:51:58 +03:00
96cba45d71
Modify xformers instead of pytorch
2023-05-18 17:29:47 +08:00
182330ae40
Merge branch 'dev' into ngrok-py
2023-05-18 10:12:17 +03:00
0d31f20cbd
Use ngrok-py library
2023-05-16 16:09:41 -04:00
9a9557ecfc
Change to extra-index-url
2023-05-15 03:00:23 +08:00
d9968e6108
launch.py: Don't involve shell for running Python or Git for output
...
Fixes Linux regression in 451d255b58
2023-05-14 20:39:19 +03:00
f29c41bf6d
Modify pytorch command
2023-05-14 22:29:28 +08:00
451d255b58
Get rid of check_run + run_python
2023-05-12 20:54:06 +03:00
55d222a9f4
launch.py: make git_tag() and commit_hash() work even when WEBUI_LAUNCH_LIVE_OUTPUT
2023-05-12 20:54:06 +03:00
681c16dd1e
fix --data-dir for COMMANDLINE_ARGS
...
move reading of COMMANDLINE_ARGS into paths_internal.py so --data-dir can be properly read
2023-05-12 22:33:21 +09:00
49a55b410b
Autofix Ruff W (not W605) (mostly whitespace)
2023-05-11 20:29:11 +03:00
dd3ca9adf7
launch.py: make torch_index_url an envvar
2023-05-11 11:57:46 +03:00
a09e1e6e18
launch.py: Use GitHub archive URLs for gfpgan, clip, openclip instead of git clones
2023-05-11 11:57:43 +03:00
875bc27009
launch.py: Simplify run()
2023-05-11 11:57:41 +03:00
49db24ce27
launch.py: Add debugging envvar to see install output
2023-05-11 11:57:36 +03:00
f5ea1e9d92
bump torch version
2023-05-10 07:26:42 +03:00
ab4ab4e595
add version to infotext, footer and console output when starting
2023-05-08 15:23:49 +03:00
5e4a0e3d24
attempt to fix broken github CI
2023-04-29 23:02:23 +03:00
97167a5768
Merge pull request #9165 from Ming424/update-readme
...
Small update for readme
2023-04-29 17:39:02 +03:00
1bab1797c0
use parsed commandline args for --skip-install
2023-04-29 17:07:21 +03:00
ce64cab397
Merge branch 'dev' into patch-1
2023-04-29 17:04:37 +03:00
ee71eee181
stuff related to torch version change
2023-04-29 12:36:50 +03:00
7fb72edaff
change index url
2023-04-13 06:47:48 -04:00
80752f43b2
revert xformers
2023-04-04 17:27:27 -04:00
4fa59b045a
update xformers
2023-04-03 15:23:35 -04:00
d537a1f1b6
Fix skip-install bug (see #8935 )
2023-04-04 00:14:20 +09:00
d5063e07e8
update torch
2023-03-30 10:57:54 -04:00
384bfe22cd
Update launch.py
2023-03-29 17:00:20 -04:00
daa5a83bb7
Merge pull request #9052 from space-nuko/temp-disable-extensions
...
Temporary disable extensions option
2023-03-28 19:42:09 +03:00
f1db987e6a
Merge pull request #8958 from MrCheeze/variations-model
...
Add support for the unclip (Variations) models, unclip-h and unclip-l
2023-03-28 19:39:20 +03:00
56f62d3851
Skip extension installers if all disabled
2023-03-27 17:23:20 -04:00
fc8e1008ea
Make disable configurable between builtin/extra extensions
2023-03-27 12:44:49 -04:00
2a4d3d2124
Add temporary "disable all extensions" option for debugging use
2023-03-27 12:04:45 -04:00
8a34671fe9
Add support for the Variations models (unclip-h and unclip-l)
2023-03-25 21:03:07 -04:00
8c801362b4
split commandline args into its own file
...
make launch.py use the same command line argument parser as the main program
2023-03-25 16:05:25 +03:00
c1294d849a
make it possible for user to enable gradio analytics by setting GRADIO_ANALYTICS_ENABLED=True
2023-03-25 12:21:18 +03:00
501f40d834
Merge pull request #8658 from hananbeer/disable_gradio_analytics
...
preserve privacy by disabling gradio analytics globally
2023-03-25 12:19:40 +03:00
layerdiffusion