cmdr2
7982a9ae25
Change the performance field to GPU Memory Usage instead, and use the 'balanced' profile by default, since it's just 5% slower than 'high', and uses nearly 50% less VRAM
2022-12-16 11:34:49 +05:30
cmdr2
aa01fd058e
Set performance level (low, medium, high) instead of a Turbo field. The previous Turbo field is equivalent to 'Medium' performance now
2022-12-15 23:30:06 +05:30
cmdr2
35ff4f439e
Refactor save_to_disk
2022-12-14 16:30:19 +05:30
cmdr2
d103693811
Bug in the metadata generation - made an array of None
2022-12-14 10:22:24 +05:30
cmdr2
6cd0b530c5
Simplify the code for VAE loading, and make it faster to load VAEs (because we don't reload the entire SD model each time a VAE changes); Record the error and end the thread if the SD model fails to load during startup
2022-12-13 15:46:04 +05:30
cmdr2
a483bd0800
No need to catch and report exceptions separately in the renderer now
2022-12-13 11:46:13 +05:30
cmdr2
47a39569bc
Merge branch 'beta' into refactor
2022-12-13 11:45:43 +05:30
cmdr2
f00e1a92d8
Don't hang the task if something other than the renderer fails (e.g. model loading)
2022-12-13 11:44:20 +05:30
Jens Sülwald
b750c0d7c3
The exception should also mention dpm2
2022-12-12 16:24:03 +01:00
cmdr2
a244a6873a
Use the new 'diffusionkit' package name
2022-12-12 20:46:11 +05:30
cmdr2
27963decc9
Use the multi-filters API
2022-12-12 18:12:55 +05:30
cmdr2
07bd580050
Typos
2022-12-12 15:44:22 +05:30
cmdr2
ac0961d7d4
Typos from the refactor
2022-12-12 15:18:56 +05:30
cmdr2
4bbf683d15
Minor refactor
2022-12-12 14:41:36 +05:30
cmdr2
b57649828d
Refactor the save-to-disk code, moving parts of it to diffusionkit
2022-12-12 14:01:47 +05:30
cmdr2
e45cbbf1ca
Use the turbo setting if requested
2022-12-11 20:42:31 +05:30
cmdr2
1a5b6ef260
Rename runtime2.py to renderer.py; Will remove the old runtime soon
2022-12-11 20:21:25 +05:30
cmdr2
096556d8c9
Move away the remaining model-related code to the model_manager
2022-12-11 20:13:44 +05:30
cmdr2
97919c7e87
Simplify the runtime code
2022-12-11 19:58:12 +05:30
cmdr2
0aa7968503
Move color correction to diffusionkit; Rename color correction to 'Preserve color profile'
2022-12-11 19:34:07 +05:30
cmdr2
6ce6dc3ff6
Get rid of the ugly copying around (and maintaining) of multiple request-related fields. Split into two objects: task-related fields, and render-related fields. Also remove the ability for request-defined full-precision. Full-precision can now be forced by using a USE_FULL_PRECISION environment variable
2022-12-11 18:16:29 +05:30
cmdr2
d03eed3859
Simplify the logic for reloading gfpgan and realesrgan models (based on the request), using the code path used for the other model types
2022-12-11 14:14:59 +05:30
cmdr2
afb88616d8
Load the models after the device init, to let the UI load before the models finish loading
2022-12-11 13:30:16 +05:30
cmdr2
543f13f9a3
Tweak logging to increase the space available by 3 characters
2022-12-11 13:19:22 +05:30
cmdr2
a2af811ad2
Disable uvicorn access logging in favor of cleaner server-side logging, we already get all that info; Print the request metadata
2022-12-09 22:47:34 +05:30
cmdr2
cde8c2d3bd
Use a logger
2022-12-09 21:30:18 +05:30
cmdr2
79cc84b611
Option to apply color correction (balances the histogram) during inpainting; Refactor the runtime to use a general-purpose dict
2022-12-09 19:39:56 +05:30
cmdr2
f1de0be679
Fix integration issues after the refactor
2022-12-09 17:50:33 +05:30
cmdr2
dbac2655f5
Typo
2022-12-09 16:14:04 +05:30
cmdr2
3fbb3f6773
Use const
2022-12-09 16:09:10 +05:30
cmdr2
8820814002
Simplify the API for resolving model paths; Code cleanup
2022-12-09 15:45:36 +05:30
cmdr2
b40fb3a422
Model readme file write flag
2022-12-09 15:27:40 +05:30
cmdr2
aa59575df3
Remove unused patch files
2022-12-09 15:24:55 +05:30
cmdr2
16410d90b8
Use the simplified model loading API in diffusion-kit; Catch and report exceptions while generating images
2022-12-09 15:21:49 +05:30
cmdr2
27c6113287
Support hypernetworks; moves the hypernetwork module to diffusion-kit
2022-12-09 13:29:06 +05:30
cmdr2
f4a6910ab4
Work-in-progress: refactored the end-to-end codebase. Missing: hypernetworks, turbo config, and SD 2. Not tested yet
2022-12-08 21:39:09 +05:30
cmdr2
bad89160cc
Work-in-progress model loading
2022-12-08 13:50:46 +05:30
cmdr2
5782966d63
Merge branch 'beta' into refactor
2022-12-08 11:58:09 +05:30
Marc-Andre Ferland
ba2c966329
First draft of multi-task in a single session. ( #622 )
2022-12-08 11:12:46 +05:30
cmdr2
fb6a7e04f5
Work-in-progress refactor of the backend, to move most of the logic to diffusion-kit and keeping this as a UI around that engine. Does not work yet.
2022-12-07 22:15:35 +05:30
cmdr2
bfdf487d52
SD2 models no longer need to be prefixed with 'sd2_' . The model loader now checks for a key that only SD2 models seem to have, to deduce which config file to use
2022-12-07 16:19:46 +05:30
Marc-Andre Ferland
182236e742
Hypernets mergefixes ( #625 )
...
* Add hypernetwork args definition in the engine.
* Add the values to reqBody
* Don't load hypernetwork.py with SD2 until it's compatible.
2022-12-07 12:35:36 +05:30
Guillaume Mercier
cbe91251ac
Hypernetwork support ( #619 )
...
* Update README.md
* Update README.md
* Make on_sd_start.sh executable
* Merge pull request #542 from patriceac/patch-1
Fix restoration of model and VAE
* Merge pull request #541 from patriceac/patch-2
Fix restoration of parallel output setting
* Hypernetwork support
Adds support for hypernetworks. Hypernetworks are stored in /models/hypernetworks
* forgot to remove unused code
Co-authored-by: cmdr2 <secondary.cmdr2@gmail.com>
2022-12-07 11:24:16 +05:30
JeLuF
e7ca8090fd
Make JPEG Output quality user controllable ( #607 )
...
Add a slider to the image options for the JPEG quality
For PNG images, the slider is hidden.
2022-12-05 11:02:33 +05:30
JeLuF
7861c57317
Safetensor support ( Fixes #599 ) ( #608 )
...
* safetensors support
Add support for checkpoints in safetensors format: https://github.com/huggingface/safetensors
This format shall be safer than pickle files
* pip install safetensors
2022-12-05 10:59:48 +05:30
Marc-Andre Ferland
367e7f7065
Add dpm2 ( #592 )
...
* Move cond_stage_model to the right device
* Removed unused vars.
* Added 'dpm2'
2022-12-02 12:58:00 +05:30
Marc-Andre Ferland
ee19eaae62
Fix for RuntimeError, missing lines. ( #591 )
...
* Move cond_stage_model to the right device
* Removed unused vars.
2022-12-02 12:57:26 +05:30
JeLuF
427861cf13
Add Quadro T2000 to force_full_precision list.
2022-12-01 00:59:12 +01:00
cmdr2
426f92595e
Merge pull request #520 from madrang/fix-gfpgan
...
Fix the gfpgan fix for multi-gpu
2022-11-30 13:08:10 +05:30
cmdr2
ac605e9352
Typos and minor fixes for sd 2
2022-11-29 13:30:08 +05:30
