JeLuF
b99d9db8f9
Create exactly 'total' images even if 'in parallel' is no factor of 'total'
2023-02-05 17:09:56 +01:00
cmdr2
368967fbcf
Merge branch 'beta' of github.com:cmdr2/stable-diffusion-ui into beta
2023-02-03 21:41:23 +05:30
cmdr2
a9d0fc9978
changelog
2023-02-03 21:41:12 +05:30
cmdr2
b6f3d2ec02
Formatting
2023-02-03 21:40:08 +05:30
cmdr2
78e917a6fb
Fix the broken 'Make Similar Images' button
2023-02-03 21:40:03 +05:30
cmdr2
96b45385e8
Merge pull request #803 from JeLuF/patch-10
...
Add T600 to list of FP only GPUs
2023-02-03 19:56:54 +05:30
cmdr2
db47888a75
changelog
2023-02-01 11:54:05 +05:30
cmdr2
51443741b8
Proactively delete the partial samples from the callbacks
2023-02-01 11:50:50 +05:30
cmdr2
3e7f14af2c
Don't use Rich Tracebacks, can cause a memory leak. It keeps a reference to the Exception object (which in turn keeps references to any torch Tensors in the stack, preventing their garbage-collection)
2023-02-01 11:50:27 +05:30
cmdr2
733439da07
Fix a memory leak. Apparently the Exception object keeps references to torch Tensors in the stack, so keeping a reference to the Exception object prevents those Tensors from getting garbage-collected.
2023-02-01 11:49:18 +05:30
JeLuF
efba81cb66
Add T1000, make Quadro equivalent to nvidia or geforce
2023-01-28 20:51:01 +01:00
JeLuF
b2cc5dcf4b
Add T600 to list of FP only GPUs
...
https://discord.com/channels/1014774730907209781/1068948110304354314
2023-01-28 20:18:07 +01:00
cmdr2
fab86ddf35
changelog
2023-01-27 09:46:50 +05:30
cmdr2
f3a90ce02d
Formatting tweaks and tip about merging similar type of models
2023-01-25 20:05:27 +05:30
cmdr2
4886616c48
changelog
2023-01-25 19:52:28 +05:30
cmdr2
dcd8121009
Revert "Temporarily disable the Merge Models UI"
...
This reverts commit 59adaf6225
2023-01-25 19:51:12 +05:30
cmdr2
59adaf6225
Temporarily disable the Merge Models UI
2023-01-25 19:46:55 +05:30
cmdr2
0055cd9b2e
Merge pull request #734 from JeLuF/mrguipi
...
Frontend of the batch merger
2023-01-25 19:39:19 +05:30
cmdr2
fe89d487f6
Merge pull request #733 from JeLuF/mrgui
...
Backend side merge API
2023-01-25 19:38:21 +05:30
patriceac
01368ac496
Add support for Windows path names
2023-01-25 02:47:50 -08:00
cmdr2
495064985e
Reduce VRAM usage of img2img in balanced mode, without reducing the speed of rendering
2023-01-24 18:58:15 +05:30
patriceac
200f8fd245
Code cleanup
2023-01-24 01:53:22 -08:00
patriceac
64bf4356b4
Update save_utils.py
2023-01-24 01:48:16 -08:00
patriceac
8d4d409cd6
Add 'embed' and 'none' to metadata saving options
...
*** Please merge https://github.com/easydiffusion/sdkit/pull/9 before merging this one. ***
This is the ED client part of metadata embedding. It adds 'embed' and 'none' options to the metadata setting and makes none the default (if never set before) because (1) it feels weird to create metadata files by default and (2) embedding by default could cause be problematic if users don't realize it's happening.
Also fixes the disabling of the dropdown in the settings when Save images to disk is toggled off.
2023-01-24 01:47:48 -08:00
patriceac
dd4937178f
Fix the tooltip display over image modifier cards
2023-01-24 01:37:37 -08:00
cmdr2
e12387a377
changelog
2023-01-23 21:40:50 +05:30
cmdr2
5d3fb9091a
Reduce the VRAM usage for balanced mode, without sacrificing the rendering speed
2023-01-23 19:36:00 +05:30
JeLuF
b044bc1791
Support multiple GFPGAN models
...
Add scanning for models and a dropdown to choose different models from
2023-01-19 20:49:54 +01:00
JeLuF
409ec61be2
Fail fatally, add same check on Linux, add some extra checks on Linux
...
Linux: Check that curl, bzip2, tar are available, check whether there's a space character in the install path, check whether the CPU supports AVX.
2023-01-19 00:40:20 +01:00
cmdr2
e2ae2715a3
Revert "Revert "Don't set the specific vram optimizations to use, instead use the new sdkit API for setting the vram usage level directly""
...
This reverts commit 52458ae273
2023-01-18 17:03:14 +05:30
cmdr2
52458ae273
Revert "Don't set the specific vram optimizations to use, instead use the new sdkit API for setting the vram usage level directly"
...
This reverts commit 42f9abdfe3
2023-01-18 10:30:56 +05:30
JeLuF
79d112ca7b
Warn when running installer from git checkout
2023-01-18 00:11:23 +01:00
cmdr2
9b1a9cc7c8
changelog
2023-01-17 21:34:41 +05:30
cmdr2
42f9abdfe3
Don't set the specific vram optimizations to use, instead use the new sdkit API for setting the vram usage level directly
2023-01-17 21:33:15 +05:30
patriceac
66d311258a
Fix card names and toggling
...
Fix names and toggling for cards starting with "By ", e.g. "By the ocean".
2023-01-16 23:56:44 -08:00
cmdr2
0a1197055c
changelog
2023-01-16 18:32:09 +05:30
cmdr2
649cbf07e3
Merge branch 'beta' of github.com:cmdr2/stable-diffusion-ui into beta
2023-01-16 18:30:46 +05:30
cmdr2
5089ac5ad1
Fix a bug where the .vae.pt extension wouldn't get picked up. Thanks Madrang, rbertus2000 and JeLuf
2023-01-16 18:30:22 +05:30
cmdr2
d99e3f7974
Merge pull request #776 from JeLuF/patch-8
...
Add NVIDIA T1200 to the list of FP GPUs
2023-01-16 18:09:06 +05:30
patriceac
3d5133209b
Group image containers (DOM tweak)
...
Move image containers in their own container to create a clear delineation in the DOM. Purely DOM structural adjustment meant to make a sticky footer possible in the preview pane (for plugins).
2023-01-15 23:34:56 -08:00
JeLuF
b5d1912c94
Add NVIDIA T1200 to the list of FP GPUs
...
Fixes https://discord.com/channels/1014774730907209781/1014774732018683926/1064269949339697163
2023-01-16 00:42:02 +01:00
patriceac
a8fba8f3fb
Fix restoration of models with subfolders
...
In dnd.js, when models are restored in the UI, there is code that strips the path from the model file name. Now that we allow models to be hosted in subfolders, this code break the task restoration (e.g. use settings, D&D, copy/paste) because "/my models/model.ckpt" becomes "model.ckpt", which won't be found.
https://discord.com/channels/1014774730907209781/1014780368890630164/1063726724573052948
2023-01-14 23:54:09 -08:00
JeLuF
9d9fc1683a
Fix number on the "Make X images" button
...
With this change, the number of prompt variants is taken into account when computing the number of images that will be generated.
X = getPrompts().length * numOutputsTotalField.value
2023-01-13 22:05:25 +01:00
cmdr2
8ee4364065
Merge pull request #768 from rbertus2000/beta
...
bugfix for FP GPUs
2023-01-13 17:39:49 +05:30
rbertus2000
152aa7de09
bugfix for FP GPUs
2023-01-13 12:54:11 +01:00
cmdr2
85c90cbee1
Merge pull request #764 from JeLuF/patch-7
...
Add NVIDIA T550 to list of FP GPUs #755
2023-01-13 10:18:24 +05:30
JeLuF
7302927e4c
Add NVIDIA T550 to list of FP GPUs #755
...
The Nvidia T550 needs full precision to work correctly.
2023-01-12 14:16:35 +01:00
cmdr2
df3d00ef94
Merge pull request #763 from patriceac/patch-18
...
Another fix for high res images
2023-01-12 10:23:01 +05:30
patriceac
bb47835256
Another fix for high res images
...
This time to address the height.
2023-01-11 17:25:54 -08:00
cmdr2
037512ca5c
Merge branch 'beta' of github.com:cmdr2/stable-diffusion-ui into beta
2023-01-11 18:25:16 +05:30
