Sean Sube
|
8c9c99eeb5
|
fix(api): use training device when loading Real ESRGAN model (#67)
|
2023-01-21 17:55:55 -06:00 |
Sean Sube
|
5286c4f596
|
fix(build): add params file to other images
|
2023-01-21 17:55:30 -06:00 |
Sean Sube
|
505cacfbc2
|
fix(api): leave pipelines on default device unless specified
|
2023-01-21 17:04:48 -06:00 |
Sean Sube
|
52484e6e1f
|
fix(build): add cv2 deps to CUDA and ROCm images
|
2023-01-21 17:01:02 -06:00 |
Sean Sube
|
a868c8cf6b
|
feat(api): backend support for multiple GPUs in diffusion pipelines
|
2023-01-21 15:17:33 -06:00 |
Sean Sube
|
a8769a5919
|
feat(build): run convert script when container starts
|
2023-01-21 15:10:59 -06:00 |
Sean Sube
|
45a3ddc2a9
|
feat(api): add option for HuggingFace token in convert script
|
2023-01-21 15:08:19 -06:00 |
Sean Sube
|
77d68bf6ac
|
fix(build): add cv2 deps to CPU image
|
2023-01-21 14:43:36 -06:00 |
Sean Sube
|
e0834110fc
|
fix(api): add extra models to convert script
|
2023-01-21 08:52:35 -06:00 |
Sean Sube
|
be16f33151
|
fix(api): pass txt2img dimensions in correct order
|
2023-01-19 22:07:31 -06:00 |
Sean Sube
|
86fb2ae28e
|
fix(api): generate correct latents for non-square images
|
2023-01-19 22:00:58 -06:00 |
Sean Sube
|
9a10261273
|
put functions in order
|
2023-01-19 21:03:29 -06:00 |
Sean Sube
|
4a3bb97342
|
fix(api): run GC after changing pipeline (#58)
|
2023-01-19 19:46:36 -06:00 |
Sean Sube
|
9a2e7adfb8
|
fix(api): report accurate image size when upscaling
|
2023-01-19 19:42:27 -06:00 |
Sean Sube
|
3bcd7a8156
|
feat: add ROCm provider to list (#10)
|
2023-01-19 17:38:26 -06:00 |
Sean Sube
|
b18567ca43
|
feat(build): add DirectML and ROCm images
|
2023-01-19 17:16:59 -06:00 |
Sean Sube
|
4fbf046e28
|
lint(api): remove unnecessary pass
|
2023-01-19 08:34:38 -06:00 |
Sean Sube
|
3679735d86
|
feat: add fill color control to inpaint
|
2023-01-18 08:41:02 -06:00 |
Sean Sube
|
08dbc0c738
|
fix(api): check image size before blending
|
2023-01-18 08:29:55 -06:00 |
Sean Sube
|
be3a17b2ff
|
feat: add version check to parameters
|
2023-01-17 23:06:04 -06:00 |
Sean Sube
|
4d9dc05cd0
|
lint(api): improve logging while upscaling
|
2023-01-17 22:30:25 -06:00 |
Sean Sube
|
cf6a151548
|
fix(api): pass image size to upscale job
|
2023-01-17 21:53:44 -06:00 |
Sean Sube
|
634d2e8ce6
|
fix(api): use correct base path for debug images
|
2023-01-17 17:50:56 -06:00 |
Sean Sube
|
5ba752e526
|
feat(api): add strength param to inpaint, remove same from upscale
|
2023-01-17 17:50:36 -06:00 |
Sean Sube
|
b496e7121c
|
fix(api): premultiply noise before compositing
|
2023-01-17 17:49:46 -06:00 |
Sean Sube
|
4aeee60b19
|
feat: add upscaling tab and endpoint
|
2023-01-16 23:45:54 -06:00 |
Sean Sube
|
b7c85aa51b
|
fix(api): skip upscaling if scale is 1
|
2023-01-16 23:01:15 -06:00 |
Sean Sube
|
227056d976
|
fix(api): correct denoise max, add missing face strength param
|
2023-01-16 22:55:54 -06:00 |
Sean Sube
|
0e27cc830d
|
fix: send missing model params, add face strength control
|
2023-01-16 22:54:01 -06:00 |
Sean Sube
|
ac3ef0ebbf
|
mostly fix face correction
|
2023-01-16 22:35:34 -06:00 |
Sean Sube
|
c6ce43b00d
|
back to single GFPGAN, fix some paths
|
2023-01-16 21:55:54 -06:00 |
Sean Sube
|
03e06193eb
|
attempt to fix face correction at various scales
|
2023-01-16 21:36:50 -06:00 |
Sean Sube
|
073ff8e02f
|
fix(api): use correct scale for background correction
|
2023-01-16 21:12:17 -06:00 |
Sean Sube
|
b09feda474
|
fix(api): sort models without discarding
|
2023-01-16 20:58:08 -06:00 |
Sean Sube
|
353a65513f
|
fix(api): copy checkpoints into correct location, handle more models
|
2023-01-16 20:53:12 -06:00 |
Sean Sube
|
1db5ebec84
|
syntax fix
|
2023-01-16 20:40:01 -06:00 |
Sean Sube
|
d52c22e58b
|
feat: add additional Real ESRGAN models, strings for them
|
2023-01-16 20:38:16 -06:00 |
Sean Sube
|
33eebacc0c
|
lint(api): return models in order
|
2023-01-16 20:37:59 -06:00 |
Sean Sube
|
c0ca7cf62f
|
fix(api): dedupe models after removing extension
|
2023-01-16 20:28:29 -06:00 |
Sean Sube
|
4472a6fd24
|
fix(api): trim model names relative to model path
|
2023-01-16 20:25:09 -06:00 |
Sean Sube
|
23a9d5afb7
|
list models correctly
|
2023-01-16 20:17:01 -06:00 |
Sean Sube
|
0273dea2a6
|
fix(api): join globs to avoid py 3.10-only args
|
2023-01-16 20:14:11 -06:00 |
Sean Sube
|
ee6308a091
|
feat(api): return all types of models
|
2023-01-16 20:10:52 -06:00 |
Sean Sube
|
dba6113c09
|
fix(api): include model scale
|
2023-01-16 20:10:29 -06:00 |
Sean Sube
|
556d5b84d6
|
fix(api): add an option to skip certain models during conversion
|
2023-01-16 17:55:58 -06:00 |
Sean Sube
|
dab2cc7752
|
lint(api): split up face correction and upscaling models
|
2023-01-16 17:51:32 -06:00 |
Sean Sube
|
decb2813c6
|
feat(api): add support for Stable Diffusion models to conversion script
|
2023-01-16 17:48:50 -06:00 |
Sean Sube
|
4d0898a52c
|
feat(api): start adding model sources to convert script
|
2023-01-16 17:08:59 -06:00 |
Sean Sube
|
5a01fe4cb0
|
fix(api): resolve face correction model relative to model path
|
2023-01-16 16:45:17 -06:00 |
Sean Sube
|
1d90f9b53c
|
fix class method
|
2023-01-16 16:40:59 -06:00 |