Commit Graph

365 Commits

Author SHA1 Message Date
henryruhs
a58a376b87 Run test on Linux and Windows 2023-06-05 15:52:33 +02:00
henryruhs
7f8406588e Run test on Linux and Windows 2023-06-05 15:49:32 +02:00
henryruhs
7426295471 Run test on Linux and Windows 2023-06-05 15:47:02 +02:00
henryruhs
95096442e2 Run test on Linux and Windows 2023-06-05 15:43:49 +02:00
henryruhs
49d8103a53 Run test on Linux and Windows 2023-06-05 15:22:22 +02:00
henryruhs
d8c6581900 Remove args overrides 2023-06-05 15:12:22 +02:00
henryruhs
249e6b3923 Remove path normalization 2023-06-05 15:01:26 +02:00
henryruhs
4420ba5bdd Normalize ARGS 2023-06-05 14:45:33 +02:00
henryruhs
b116f2001a Adjust comment 2023-06-05 12:42:24 +02:00
Henry Ruhs
3747fc2301
Merge pull request #338 from phineas-pta/next
ffmpeg platform-agnostic hardware-acceleration
2023-06-05 12:39:05 +02:00
Phan Tuấn Anh
71af0fac69
syntax error 2023-06-05 12:29:30 +02:00
Phan Tuấn Anh
b947603441
torch check nvidia only 2023-06-05 12:25:35 +02:00
Phan Tuấn Anh
3668360520
check torch gpu before clearing cache 2023-06-05 12:21:37 +02:00
henryruhs
30713c3c0e Hotfix 2023-06-05 12:10:23 +02:00
Phan Tuấn Anh
c59400f6a6
clear CUDA cache after swapping
on low VRAM + ffmpeg cuda acceleration, clearing cache prevent cuda out-of-memory error
2023-06-05 11:58:24 +02:00
Somdev Sangwan
e6cc97daab
add henryruhs to credits 2023-06-05 15:28:04 +05:30
Phan Tuấn Anh
6daee96029
ffmpeg platform-agnostic hardware-acceleration 2023-06-05 11:57:11 +02:00
Henry Ruhs
be0b57e47a
Merge pull request #331 from s0md3v/next
Next
2023-06-05 09:30:11 +02:00
henryruhs
6b0f3f93c8 Fix memory leak thanks to tensorflow 2023-06-04 22:22:35 +02:00
henryruhs
9c88540629 Better print exceptions, Change defaults 2023-06-04 21:31:07 +02:00
henryruhs
86e4eed212 It needs 1 threads or cores of course 2023-06-04 20:05:17 +02:00
Henry Ruhs
01ab68708d
Merge pull request #324 from s0md3v/next-polishing
Next polishing
2023-06-04 19:08:14 +02:00
henryruhs
b001fb5a89 Doing args the proper way 2023-06-04 18:37:55 +02:00
henryruhs
9eaeecaab5 Skip cpu pool for one cpu core to be used 2023-06-04 17:58:47 +02:00
henryruhs
5f3e30fadd Cleanup a bit and revert removal of ENV 2023-06-04 17:28:46 +02:00
henryruhs
868fa1a00d Rename max-cores to cpu-cores, Improve dependencies 2023-06-04 17:08:12 +02:00
Somdev Sangwan
08594cde35 fix multiprocessing 2023-06-04 17:40:43 +05:30
Pikachu~~~
f200b4c7b4
Roop-multi changed the implementation of multi-threading processing for nvidia GPU. (#317)
* changed the multi-thread implementation for nvidia gpu

* Update requirements.txt

* Add files via upload

* fix core.py and swapper.py

* fix core.py

* code clean

* code clean

* doubles performance of gpu-mode

---------

Co-authored-by: Moeblack <Moeblack@kuroinekorachi@gmail.com>
Co-authored-by: Somdev Sangwan <s0md3v@gmail.com>
2023-06-04 17:19:27 +05:30
Somdev Sangwan
160a16f4b5
revert changes 2023-06-04 17:18:18 +05:30
Somdev Sangwan
d05368c361
revert changes 2023-06-04 17:17:19 +05:30
Somdev Sangwan
c36d3a2009
fix multiprocessing 2023-06-04 17:00:57 +05:30
Somdev Sangwan
c19f8125f2
add threading 2023-06-04 17:00:09 +05:30
henryruhs
81320ad5f1 Fix CPU only mode 2023-06-04 10:53:01 +02:00
Somdev Sangwan
8dbbaf0bdb
clean pre_check 2023-06-04 14:13:08 +05:30
Somdev Sangwan
102f783a2d
fix providers for non-nvidia 2023-06-04 14:03:55 +05:30
Somdev Sangwan
2606be04ba
add pynvml 2023-06-04 13:24:32 +05:30
henryruhs
289bcff9a3 CoreMLExecutionProvider support for Apple Silicon 2023-06-04 09:51:05 +02:00
Somdev Sangwan
90f1c39dd5
Merge pull request #300 from K1llMan/next
Preview window
2023-06-04 13:09:56 +05:30
K1llM@n
0d624e8a9b Fix frames size 2023-06-03 17:57:40 +03:00
K1llM@n
4e5dceef6d Merge branch 'next' of https://github.com/K1llMan/roop into next 2023-06-03 17:51:49 +03:00
K1llM@n
2c91270090 Functionional style ui 2023-06-03 17:47:59 +03:00
K1llM@n
6288cdce65 Move ui from core 2023-06-03 16:02:51 +03:00
henryruhs
777261c8f9 Add cuda utilization and cuda memory to progress bar 2023-06-03 14:49:27 +02:00
henryruhs
21e2953420 Double GPU performance 2023-06-03 13:43:56 +02:00
henryruhs
1f0c3b4bb5 Double GPU performance 2023-06-03 13:37:15 +02:00
henryruhs
b0a7b19b1b Skip matrix build for now (too complicated as of cuda runtimes) 2023-06-03 12:29:51 +02:00
henryruhs
23d8d78c7e Introduce CI matrix 2023-06-03 12:08:06 +02:00
henryruhs
10433d1581 Introduce CI matrix 2023-06-03 12:01:18 +02:00
henryruhs
98e19e1523 Introduce CI matrix 2023-06-03 12:00:02 +02:00
henryruhs
9101f8d5f2 Introduce CI matrix 2023-06-03 11:52:03 +02:00