guys does any of you have upscaler workflow for a ready video
guys does any of you have upscaler workflow for a ready video
Fyou can simply extract every frame
Fthen upscale
Ftakes long time when not multithreaded
BDepending on my settings, I frequently see this kind of pattern in clothing, close up and far away, has anyone seen it before / know what it may be from?
Balmost like a snakeskin pattern lol
F100% result of LoRA
CHi guys, I'm new here, I'm trying to make a swap-face video using the google colab option (because i'm not a lot into programing) and it doesn't work. It shows a lot of error messages that I would be glad to share with you.
Can you provide some help in this regard?
Can you provide some help in this regard?
Fignor all errors
Fi just tested yesterday evening
Fworking great
F
RHey, im trying to follow the kaggle tutorial, but when I run the final cell with the
webui.sh, it runs for a second, then errors with ERROR: Operation cancelled by user, and then the whole notebook turns off, and I need to start again. Any advice?
Fkaggle banned free share
Fbut we have updated method on patreon
Fthat works even better than before
FRun Automatic1111 On A Free Kaggle NoteBook Like In Your PC - Supports SDXL & ControlNet
FKohya SD 1.5 & SDXL LoRA - DreamBooth Training Free Kaggle NoteBook
RThanks. Just became a member
Fawesome welcome
Fyou will also see mini tutorial inside post
RThanks for all our awesome tutorials!
Fty
UHi. I'm getting this error trying to train voice with DL-Art-School:
Loading from ../experiments/dvae.pth
Traceback (most recent call last):
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\train.py", line 398, in <module>
trainer.init(args.opt, opt, args.launcher)
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\train.py", line 146, in init
self.model = ExtensibleTrainer(opt)
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\trainer\ExtensibleTrainer.py", line 105, in init
step = ConfigurableStep(step, self.env)
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\trainer\steps.py", line 49, in init
self.injectors.append(create_injector(injector, env))
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\trainer\inject.py", line 65, in create_injector
return injectors[opt_inject['type']](opt_inject, env)
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\trainer\injectors\audio_injectors.py", line 178, in init
self.dvae = load_model_from_config(cfg, dvae_name, device=f'cuda:{env["device"]}').eval()
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\utils\util.py", line 497, in load_model_from_config
sd = torch.load(load_path, map_location=device)
File "C:\Users\hasan\miniconda3\envs\DLAS\lib\site-packages\torch\serialization.py", line 777, in load
with _open_zipfile_reader(opened_file) as opened_zipfile:
File "C:\Users\hasan\miniconda3\envs\DLAS\lib\site-packages\torch\serialization.py", line 282, in init
super(_open_zipfile_reader, self).init(torch._C.PyTorchFileReader(name_or_buffer))
RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory
Press any key to continue . . .
Loading from ../experiments/dvae.pth
Traceback (most recent call last):
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\train.py", line 398, in <module>
trainer.init(args.opt, opt, args.launcher)
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\train.py", line 146, in init
self.model = ExtensibleTrainer(opt)
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\trainer\ExtensibleTrainer.py", line 105, in init
step = ConfigurableStep(step, self.env)
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\trainer\steps.py", line 49, in init
self.injectors.append(create_injector(injector, env))
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\trainer\inject.py", line 65, in create_injector
return injectors[opt_inject['type']](opt_inject, env)
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\trainer\injectors\audio_injectors.py", line 178, in init
self.dvae = load_model_from_config(cfg, dvae_name, device=f'cuda:{env["device"]}').eval()
File "C:\Users\hasan\voice-clone\DL-Art-School\codes\utils\util.py", line 497, in load_model_from_config
sd = torch.load(load_path, map_location=device)
File "C:\Users\hasan\miniconda3\envs\DLAS\lib\site-packages\torch\serialization.py", line 777, in load
with _open_zipfile_reader(opened_file) as opened_zipfile:
File "C:\Users\hasan\miniconda3\envs\DLAS\lib\site-packages\torch\serialization.py", line 282, in init
super(_open_zipfile_reader, self).init(torch._C.PyTorchFileReader(name_or_buffer))
RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory
Press any key to continue . . .
Fwhat is your gpu
Falso
Funinstall all python
Finstall python 3.10.x into C drive
Finstall voice clan into C drive
Fnot in users folder
UI think there was some files that are not downloaded correctly. When I delete everything and install again, it worked.
WSir, does VGA/GPU affect the quality of deepfake?
Fnope just speed
Ihas anybody done video with gpt4v? can we? extract image, combine RAG or something?
can we somehow use dall e to create video? create sequence of image and 'stictch' them together?
can we somehow use dall e to create video? create sequence of image and 'stictch' them together?
Fi doubt that you can do
Iwhy not? thanks for the response
Fit can generetate individual images
Fnot consisten to make video
Ii mean video2text is possible?
workflow is: extract image out of video, image2text using GPT4V, combining text together with prompt engineering, store them in embedding/vector db, finetune a model with those embeddings, then generate such images with DALL-E 3, consume 3rd party API to stitch them together?
workflow is: extract image out of video, image2text using GPT4V, combining text together with prompt engineering, store them in embedding/vector db, finetune a model with those embeddings, then generate such images with DALL-E 3, consume 3rd party API to stitch them together?
Ii think openAI has new feature to make DALL-3 more consistent now
Iat least a slide show now
Fyes possible
Feasy
Fwith GPT4 vision API
Ithanks, any feedback on my approach? cuz vision api is mostly for images not videos, ive not tried the other way round though, whether image2text (gpt4v) and text2image (dall-3) can be made consistent, with the new feature?
Fyou need to extract