Commit Graph

72 Commits

Author SHA1 Message Date
DGdev91 62860c221e Skip force pyton and pytorch ver if TORCH_COMMAND already set 2023-06-06 15:43:32 +02:00
DGdev91 8646768801 Write "RX 5000 Series" instead of "Navi" in err 2023-06-06 10:03:20 +02:00
DGdev91 95d4d650d4 Check python version for Navi 1 only 2023-06-06 09:59:13 +02:00
DGdev91 e0d923bdf8 Force python1 for Navi1 only, use python_cmd for python 2023-06-06 09:55:49 +02:00
DGdev91 2788ce8c7b Fix error in webui.sh 2023-06-06 01:51:35 +02:00
DGdev91 8d98532b65 Forcing Torch Version to 1.13.1 for Navi and Renoir GPUs 2023-06-06 01:05:31 +02:00
Aarni Koskela 46a5bd64ed Restart: only do restart if running via the wrapper script 2023-06-05 20:04:28 +03:00
AUTOMATIC 333e63c091 a yet another method to restart webui 2023-06-03 09:59:56 +03:00
AUTOMATIC1111 7fd53815d3
Merge pull request #10967 from waltercool/master
Added support for workarounds on Navi external GPU.
2023-06-03 09:09:25 +03:00
Pablo Cholaky 8d970a4a97
Added support for workarounds on external GPU.
lspci detects VGA for main/integrated videocards and Display
for external videocards.

This commit should apply workarounds on computers with more than
one GPU. Useful for most laptops using weak iGPU and good dGPU.

Signed-off-by: Pablo Cholaky <waltercool@slash.cl>
2023-06-02 15:04:58 -04:00
AUTOMATIC1111 348abeb99d
Merge pull request #10722 from maybe-hello-world/master
Download ROCm for AMD GPU only if NVIDIA is not presented
2023-05-27 19:56:18 +03:00
AUTOMATIC1111 ba812b4495
Merge pull request #10718 from kernelmethod/libtcmalloc_fixes
Small fixes to prepare_tcmalloc for Debian/Ubuntu compatibility
2023-05-27 19:55:02 +03:00
Roman Beltiukov bdc371983e
Update webui.sh 2023-05-26 02:09:09 -07:00
Roman Beltiukov b2530c965c
Merge branch 'dev' into master 2023-05-25 15:10:10 -07:00
Roman Beltiukov 09d9c3d287
change to AMD only if NVIDIA is not presented 2023-05-25 14:45:05 -07:00
kernelmethod d29fe44e46 Small fixes to prepare_tcmalloc for Debian/Ubuntu compatibility
- /usr/sbin (where ldconfig is usually located) is not typically on users' PATHs by default, so we set that variable before trying to run ldconfig.
- The libtcmalloc library is called libtcmalloc_minimal on Debian/Ubuntu systems. We now check whether libtcmalloc_minimal exists when running prepare_tcmalloc.
2023-05-25 14:51:47 -04:00
AUTOMATIC1111 97e1cf69c0
Merge branch 'dev' into master 2023-05-18 10:26:35 +03:00
Weiming aa6e98e43c Error Improvement for install torch 2023-05-18 13:25:48 +08:00
Baptiste Rajaut 484948f5c0
Fixing webui.sh
If only i proofread what i wrote
2023-05-17 11:10:57 +02:00
Baptiste Rajaut b3397c2492
Bump pytorch for AMD Users
So apparently it works now? Before you would get "Pytorch cant use the GPU" but not anymore.
2023-05-17 11:01:33 +02:00
AUTOMATIC1111 335428c2c8
Merge pull request #9140 from yedpodtrzitko/yed/reuse-existing-venv
feat: use existing virtualenv if already active
2023-05-02 11:05:00 +03:00
AUTOMATIC 14b70aa97b revert unwanted change from #9865 2023-05-02 11:03:11 +03:00
catalpaaa 9eb5b3e90f
Merge branch 'experimental' into subpath-support 2023-05-01 11:59:21 -07:00
AUTOMATIC1111 1185bf3981
Merge branch 'dev' into master 2023-04-29 17:16:52 +03:00
AUTOMATIC1111 967fb51df2
Merge branch 'dev' into master 2023-04-29 14:09:45 +03:00
AUTOMATIC1111 1ffb44b0b2
Merge pull request #9593 from gakada/tcmalloc
Try using TCMalloc on Linux by default
2023-04-29 13:02:00 +03:00
catalpaaa b2f6e0704e add subpath support 2023-04-25 07:27:24 -07:00
weidong 152ed34ccc
fix install_dir error
When the user's home directory and username are inconsistent, an error message stating that the directory cannot be found will appear. Directly default the installation directory to the user's home directory
2023-04-17 17:17:10 +08:00
gk 8af4b3bbe4 Try using TCMalloc on Linux by default 2023-04-13 10:19:03 +09:00
DGdev91 9edd4b6e51 Using --index-url instead of --extra-index-url following new PyTorch install command 2023-04-11 11:22:28 +02:00
DGdev91 3a5b47e26e Forcing PyTorch version for AMD GPUs automatic install
The old code tries to install the newest versions of pytorch, wich is currently 2.0. Forcing it to 1.13.1
2023-04-06 01:36:27 +02:00
yedpodtrzitko 0d2cf9ac18 feat: use existing virtualenv if already active 2023-03-29 16:35:37 +07:00
ParityError 5c051c0618
Update webui.sh
Installation should not be assumed to be located within ~/home directory. User should be expected to install project anywhere and run the startup scripts while in stable-diffusion-webui directory.

See issue #8534
2023-03-12 15:10:44 -07:00
ParityError 6439e72df2
Update webui.sh
Installation should not be assumed to be located within ~/home directory. User should be expected to install project anywhere and run the startup scripts while in stable-diffusion-webui directory.

See issue #8534
2023-03-12 15:08:26 -07:00
AUTOMATIC 6033de18bf revert webui.sh from #8492 2023-03-12 20:50:02 +03:00
Zhang Hua d006108d75
webui.sh: remove all `cd` related code
This may be helpful for
https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/7028,
because we won't change working directory to the repo now, instead,
we will use any working directory. If we set working directory to
a path contains repo and the custom --data-dir, the problem in this
issue should be solved.

Howewer, this may be treated as an incompatible change if some code
assume the working directory is always the repo.

Also, there may be another solution that always let --data-dir be the
subdirectory of the repo, but personally I think this may not be what
we actually need.

As this issue mainly influent on Docker and I am not familiar with
.bat files, updating webui.bat is skipped.

webui.sh: source env from repo instead $PWD
2023-03-11 20:00:11 +08:00
DaniAndTheWeb e0b6092bc9
Update webui.sh 2023-01-20 15:31:27 +01:00
DaniAndTheWeb fd651bd0bc
Update webui.sh 2023-01-20 00:21:51 +01:00
DaniAndTheWeb 0684a6819d
Usage explanation for Renoir users 2023-01-20 00:21:05 +01:00
DaniAndTheWeb 912285ae64
Experimental support for Renoir
This adds the GFX version 9.0.0 in order to use Renoir GPUs with at least 4 GB of VRAM (it's possible to increase the virtual VRAM from the BIOS settings of some vendors). This will only work if the remaining ram is at least 12 GB to avoid the system to become unresponsive on launch.).
This change also changes the GPU check to a case statement to be able to add more GPUs efficiently.
2023-01-19 23:42:12 +01:00
DaniAndTheWeb 36364bd76c
GFX env just for RDNA 1 and 2
This commit specifies which GPUs should use the GFX variable, RDNA 3 is excluded since it uses a newer GFX version
2023-01-19 20:05:49 +01:00
DaniAndTheWeb 48045545d9
Small reformat of the GPU check 2023-01-19 19:23:40 +01:00
DaniAndTheWeb c09fb3d8f1
Simplify GPU check 2023-01-19 19:21:02 +01:00
DaniAndTheWeb 4599e8ad0a
Environment variable on launch just for Navi cards
Setting HSA_OVERRIDE_GFX_VERSION=10.3.0 for all AMD cards seems to break compatibility for polaris and vega cards so it should just be enabled on Navi
2023-01-19 17:00:51 +01:00
facu 956263b8a4 fixing error using lspci on macOsX 2023-01-18 19:15:53 -03:00
DaniAndTheWeb ba077e2110
Fix TORCH_COMMAND check 2023-01-14 23:19:52 +01:00
DaniAndTheWeb 2e172cf831
Only set TORCH_COMMAND if wasn't set webui-user 2023-01-14 22:25:32 +01:00
DaniAndTheWeb c4ba34928e
Quick format fix 2023-01-14 15:58:50 +01:00
DaniAndTheWeb 6192a222bf
Export TORCH_COMMAND for AMD from the webui 2023-01-14 15:46:23 +01:00
DaniAndTheWeb cbf4b3472b
Automatic launch argument for AMD GPUs
This commit adds a few lines to detect if the system has an AMD gpu and adds an environment variable needed for torch to recognize the gpu.
2023-01-13 19:18:56 +01:00