DeepFaceLab 2.0 Easy Tutorial | Part 1 [ 2023 ]
Вставка
- Опубліковано 20 чер 2023
- ➜ This deepfacelab tutorial for deepfacelab 2022 is for people who are thinking of deepface or deepfacelab download windows. So if you curious how does deepfacelab tutorial 2023 works for deepfacelive this video is for you. Deepfacelab pretraining an deepfake live tutorial. Thinking about making deepfake, deepfacelab hindi, deepfacelab guide this video is for you making deepfake collab. Deepfake mac or deepfacelab mac is also possible. This deepfacelab is a comprehensive guide and tutorial for deepfake.
➡NEW VIDEO: • How to Make Animated T...
THANKS FOR WATCHING GUYS! 🙏🙏🙏
Inside this video: DeepFaceLab 2.0 Easy Tutorial | Part 1 [ 2023 ]
My Social Media:
Twitter - Aiovo5
Instagram - aiovo_official
Facebook - Aiovo-107155...
Request a Sponsorship/Enquires: aiovopromo@yahoo.com (Include “Sponsor” in Email Subject) - Навчання та стиль
➡NEW VIDEO: ua-cam.com/video/BXQcaiieTGU/v-deo.html
Thank you for this very interesting tutorial. I have downloaded a pre trained model. When I run it I see that the counter of the iterations continues from that of the original pre trained model, instead in the various videos that show the difference of using a virgin model versus a pre trained one, the counter of the iterations starts from zero. What am I doing wrong? Thank you in advance for your kind support and patience!
i have a problem with the Merger , after collecting alignments the interactive merger doesn't open , its stuck on " running on Cpu 07 " the cpu number change every time i try , and the interactive merger wont launch , even if i choose not to use the interactive merger the same thing happens . anyone have an answer to that problem ? thank you
Pls I need help!. I trained my model for 48hrs over 100k. And now trying to merge SAEHD. It’s stuck on 0%. What do I do?.
The paging file is too small for this operation to complete
doesn't matter what size i make the page file, it still does this.
Do I have to delete the original source and destination video or one will just drop the custom video in the Workspace folder , I tried dropping my video but it kept showing me the initial default video with file named data_dst & data_src
I think same before version, u just rename ur resource video to data src and video set to data_dst ,src for face fake dst for background video
thanks for good tutorial :) i have a question that how can i take FPS from the data_dst ? there is no choice but too much big if i dont choice the FPS
i want to change to low FPS but same high quality of face . how can i change my setting? please help me
Choose one or several GPU idxs (separated by comma).
[CPU] : CPU
[0] : NVIDIA GeForce GTX 1050
[1] : AMD Radeon(TM) Vega 8 Graphics
[1] Which GPU indexes to choose? : 0
0
[2] Autobackup every N hour ( 0..24 ?:help ) :
2
[n] Write preview history ( y/n ?:help ) :
n
[1] Target iteration : 0
0
[n] Flip SRC faces randomly ( y/n ?:help ) :
n
[y] Flip DST faces randomly ( y/n ?:help ) :
y
[2] Batch_size ( ?:help ) :
2
[64] Resolution ( 64-640 ?:help ) :
64
[f] Face type ( h/mf/f/wf/head ?:help ) :
f
[liae-ud] AE architecture ( ?:help ) :
liae-ud
[32] AutoEncoder dimensions ( 32-1024 ?:help ) :
32
[16] Encoder dimensions ( 16-256 ?:help ) :
16
[16] Decoder dimensions ( 16-256 ?:help ) :
16
[16] Decoder mask dimensions ( 16-256 ?:help ) :
16
[n] Eyes and mouth priority ( y/n ?:help ) :
n
[n] Uniform yaw distribution of samples ( y/n ?:help ) :
n
[n] Blur out mask ( y/n ?:help ) :
n
[y] Place models and optimizer on GPU ( y/n ?:help ) :
y
[y] Use AdaBelief optimizer? ( y/n ?:help ) :
y
[n] Use learning rate dropout ( n/y/cpu ?:help ) :
n
[y] Enable random warp of samples ( y/n ?:help ) :
y
[0.0] Random hue/saturation/light intensity ( 0.0 .. 0.3 ?:help ) :
0.0
[0.0] GAN power ( 0.0 .. 5.0 ?:help ) :
0.0
[0.0] Face style power ( 0.0..100.0 ?:help ) :
0.0
[0.0] Background style power ( 0.0..100.0 ?:help ) :
0.0
[none] Color transfer for src faceset ( none/rct/lct/mkl/idt/sot ?:help ) :
none
[n] Enable gradient clipping ( y/n ?:help ) :
n
[n] Enable pretraining mode ( y/n ?:help ) :
n
the part that asks for FPS has a default value of 0. If you choose anything higher, you're telling the system to skip that many frames when pulling faces from your data_dst video. So you want this at 0 if you want it to pull faces from every frame. Hope that clears that up. Or are you experiencing another issue?
How can i save the trained face for using in another videos?! its possible?
It's awkward even if it's over a million times
There's a lot of video quality that's blurry, but if it's longer than
Do I have to do more videos?
It stopped at 0.1730
Someone told me to go from 0.02 to 0.08, is that right?
whit 4000 sx what build i need to install?
Some developer completely failed front end user interface design
You said at 7:54 that with a better GPU you get a higher speed, which is not the case. I got a RTX 4090 and it's nearly identical to your speed. Unfortunately the software is limited in this regard
This is hard to believe… likely need to do some optimization, at least save on energy since the GPU is stronger?
Where is part 2?
Can I run this on rtx 3050 4gb gpu .
Please if yes , I'm guessing batch size 2 and resolution 320 right ?
The minimum batch size is 4 since anything below will result in a not so perfect end results. I'm running on a Gtx 1660 ti with a 6gb vram. Are you sure it's only 4gb?
Hey, Can I run this on GTX 1650 with 4 GB vram?@@SmellyFist1
Hello sir. Can you create besr deep fakes on your pc of my face? I'll pay you whatever amount you sayy for that..please 🙏
shift + right arrow not working...help please? 21:50 step 7 Merge SAEHD is not working for me
same
It's not right arrow. It's " Shift + > " The key next to the Question mark and / key.
How can I use the model in my DEEPFACELIVE Model? How can I make it t be selectable on my DEEPFACELIVE model?
you need to run - 6) export SAEHD as dfm - this will put a dfm file in your model folder. copy and paste that file into the DeepFaceLive_NVIDIA - userdata - samples folder. you can then open deepfacelive and select it by going to the - Face Swap (DFM) section of DeepFaceLive and where it says model you can select it. on the face aligner section also remember to select the same resolution as you made your model.
where do you get pre trained models?
DeepFaceLive
Dude, it's 2023, why didn't you mention what the best option for a 4000 series is?
Because its the exact same as a 3000 series
Only thing i would change is from 128 to 224 resolution and run 8 batch instead of 4. If you are getting over 300it per second change back to 4 on your batch
Last but not least this is an inferior version of dfl the vanilla mve version is much better, Druuzil techs and gaming has a recent video on it and even has the files on a Google drive to download. His video is 5 hours long but it covers everything
My system seems to get stuck at loading samples for some reason? Any ideas? It's just at 100% and nothing else happens
same
@@petepablogaming243 same!
@@petepablogaming243 try diretix 12 version, you have truble with GPU
@@petepablogaming243 Turn on hardware acceleration
5:00
3:18
10:41
I'm not interested in video fakes, can DFL make fakes of singular images? None of these youtubers cover how to do that.
It makes fakes of videos by swapping a series of thousands of frames. One image would be simple.
@@smallpseudonym2844 I tried it but It didn't work too well since I only have limited number of images of pictures to work with.
PLEASE FFS STOP MOVING YOUR MOUSE!!!!! I CANT CONCENTRATE!!!
bruh canna getta part deuce tho. this shit slapped me in the easy chair and if you just teach my ass i'll hit ya with some glizzy money
4:00
10:15