MrDeepFakes Forums
  • New and improved dark forum theme!
  • Guests can now comment on videos on the tube.
   
dpfksOpenFaceSwap Installation and Use
#21
(12-25-2018, 03:01 PM)Akanee Wrote: You are not allowed to view links. Register or Login to view.But what about photos you said that it got some all in one features...but photos are not one of them right ?

The application swaps the faces onto the frames, then the frames are used to generate video.  So if you skip the generate video step, you would have several images.  But it's not included as a "feature".

So it is possible to use faking applications to get faceswapped photos, but it's really designed for video.  However, in my opinion, this is an overly-complicated way to doing something that could be done faster and easier with Photoshop or GIMP.
#22
(12-25-2018, 03:25 PM)Pocketspeed Wrote: You are not allowed to view links. Register or Login to view.
(12-25-2018, 03:01 PM)Akanee Wrote: You are not allowed to view links. Register or Login to view.But what about photos you said that it got some all in one features...but photos are not one of them right ?

The application swaps the faces onto the frames, then the frames are used to generate video.  So if you skip the generate video step, you would have several images.  But it's not included as a "feature".

So it is possible to use faking applications to get faceswapped photos, but it's really designed for video.  However, in my opinion, this is an overly-complicated way to doing something that could be done faster and easier with Photoshop or GIMP.

hum ...yeah but on photoshop it's really hard to match color seamless toning light shadows.....
Have u got any idea or "trick" to this quickly ? doing something seamless as fakeapp on videos ?
Maybe a plugin or something on photoshop ?
#23
(12-25-2018, 03:48 PM)Akanee Wrote: You are not allowed to view links. Register or Login to view.
(12-25-2018, 03:25 PM)Pocketspeed Wrote: You are not allowed to view links. Register or Login to view.
(12-25-2018, 03:01 PM)Akanee Wrote: You are not allowed to view links. Register or Login to view.But what about photos you said that it got some all in one features...but photos are not one of them right ?

The application swaps the faces onto the frames, then the frames are used to generate video.  So if you skip the generate video step, you would have several images.  But it's not included as a "feature".

So it is possible to use faking applications to get faceswapped photos, but it's really designed for video.  However, in my opinion, this is an overly-complicated way to doing something that could be done faster and easier with Photoshop or GIMP.

hum ...yeah but on photoshop it's really hard to match color seamless toning light shadows.....
Have u got any idea or "trick" to this quickly ? doing something seamless as fakeapp on videos ?
Maybe a plugin or something on photoshop ?

This isn't really the forum for that, but I used to adjust saturation, contrast, and hue to get a close match, then create and use different layers to achieve results.  There are tutorials on Youtube for this stuff.

I mean, you can use a deepfake application, with very specific requirements, for several hours to create some images.  Or you can use Photoshop or GIMP for about an hour or so.  With few specific requirements.  Faking applications automate a lot of the process, yes.  But either way, you'll be putting in some work, and you will have to learn some stuff.

The only advantage I see is that you could use deepfaking to get lots of images faster than you could manually through Photoshop.  But if it's just one or two pics you want, then Photoshop will be more efficient.  But this is just my opinion.
#24
[quote pid='1812' dateline='1545742110']
[quote pid='1810' dateline='1545740715']
The only advantage I see is that you could use deepfaking to get lots of images faster than you could manually through Photoshop.  But if it's just one or two pics you want, then Photoshop will be more efficient.  But this is just my opinion.
[/quote]

[/quote]


I think the same conerning that kind of tricks. 
Alignment etc... for photoshop still the best Smile

Thank you a lot for taken time to answer me Smile
#25
There are several patches for CUDA 9 with corresponding bugfixes and updates to cuDNN

Anyone tried updating or patching and was it worth it?



These patches require the base installer to be installed first.


Patch 1 (Released Jan 25, 2018)
cuBLAS Patch Update: This update to CUDA 9.0 includes new GEMM kernels optimized for the Volta architecture and improved heuristics to select GEMM kernels for given input sizes.

Patch 2 (Released Mar 5, 2018)
cuBLAS Patch Update: This update to CUDA 9 includes GEMM heuristics improvements to selects the most optimized algorithms for input sizes commonly used in Deep Learning RNNs. The update also includes other bug-fixes and performance enhancements.

Patch 3 (Released Jun 7, 2018)
cuBLAS Patch Update: This update to cuBLAS addresses issues with Convolutional Seq2Seq and RNN inference performance.

Patch 4 (Released Aug 6, 2018)
cuBLAS Patch Update: This update to cuBLAS includes optimized implementations of GEMV operations for mixed precision input and output types and important fixes to address performance issues.
#26
(01-20-2019, 07:17 PM)robo3d Wrote: You are not allowed to view links. Register or Login to view.There are several patches for CUDA 9 with corresponding bugfixes and updates to cuDNN

Anyone tried updating or patching and was it worth it?



These patches require the base installer to be installed first.


Patch 1 (Released Jan 25, 2018)
cuBLAS Patch Update: This update to CUDA 9.0 includes new GEMM kernels optimized for the Volta architecture and improved heuristics to select GEMM kernels for given input sizes.

Patch 2 (Released Mar 5, 2018)
cuBLAS Patch Update: This update to CUDA 9 includes GEMM heuristics improvements to selects the most optimized algorithms for input sizes commonly used in Deep Learning RNNs. The update also includes other bug-fixes and performance enhancements.

Patch 3 (Released Jun 7, 2018)
cuBLAS Patch Update: This update to cuBLAS addresses issues with Convolutional Seq2Seq and RNN inference performance.

Patch 4 (Released Aug 6, 2018)
cuBLAS Patch Update: This update to cuBLAS includes optimized implementations of GEMV operations for mixed precision input and output types and important fixes to address performance issues.

Yep.  You need the CUDA patches.  Patch 1 will just flash the command prompt window briefly, then close itself, when you install it.  The rest will be the usual installation process.  Use custom install.  Do not overwrite your drivers.  Uncheck those boxes when asked.
#27
Hello
I have an error when I press on face A


aligning faces from image A set with commande
python\scripts\python.bas faceswap\facewap.py.extract-i\imageA.\alimA-Dcn

help please
#28

J'ai une erreur quand j'appuie sur le visage A



aligner les visages de l'image un ensemble avec commande
python \ scripts \ python.bas faceswap \ facewap.py.extract-i \ imageA. \ alimA-Dcn

Aidez-moi, s'il vous plaît

Hello
I have an error when I press on face A



aligning faces from image A set with commande
python\scripts\python.bas faceswap\facewap.py.extract-i\imageA.\alimA-Dcn

help please
#29
Is it enough if i'm using my nvidia gtx850m ?
#30
Hello.. i got face detected 0.. any help ?

Forum Jump:

Users browsing this thread: 1 Guest(s)