Видео

When I run this, my resulting image looks absolutely awful. It's a bit burry and totally destroyed the face of the man in the starting image. This ControlNet must be pretty weak.

No matter what I did, the shadow under the photo does not look natural.. what should I do, what should I manipulate?

Nice ❤

not same face, a different man in upscale, the main problem with all these tools, which preventing to use for any serious restoration. It makes "eye candy" result as usual-the core script of all this Ai, but nothing related to original.

where is the controlnet model??

hi, do you know any model for comfy ui so we can create ai avatar speaker? i am searching for a comfy ui model that i give him a photo (my avatar) and a voice file , then comfy ui give me ai avatar speaker with lip+eye+face synce. i know there are websites to do it but i want to make it in my local pc for free. i have pretty good gpu.

Hey How to use the workflow? Where do I need to put the downloaded workflow?

I want to start using comfyUI. Currently I'm using swarmUI, it's backend is ComfyUI. I'm able to open it but after creating nodes and going forward it just gives error could not detect model type

just noticed, all the asset names are changed to "diffusion_pytorch_model.safetensors"

hm... for some reason it is giving my images a bit of a double vision / blur... it's not as sharp. is this a gguf thing?

Let it be know that this is a workflow for SUPER LOW Resolutuion sub 512, so this is a great help if you are upscaling a image of a scanned stamp. It will not work with even basic 832x1216, so do not get fooled.

Where to download the Fun5B models ?

Audio syncing is the only issue, can you update it as soon as possible?

what do you mean by low VRAM ? below 8 GB ? below 16 GB ?

In the CogVideoXFun Sample, make sure the base Resolution is 512

Thank you for the video. My Gym is creating samples but no Loras are made after the steps are completed. Can you offer any advice? I have 10.24 gb ram so I'm wondering if that is the issue. Thank you.

Thank you for sharing your workflow! Liked and now subscribed.

Thanks for the video. I get this error: "KSampler: mat1 and mat2 shapes cannot be multiplied (1x768 and 2816x1280)" Can you please help, what should I do? my image is 1888x1056px

Hello Where to download all these models I go to the same place you show where it says “resources” on huggingface, but there are only models “diffusion_pytorch_model.safetensors”.

Thanks a ton for this workflow and tutorial, I just need to find the right clip models and this should work for me !! It's unfortunate that I never downloaded any of the clip models used here, so I gotta look for them

Great tutorial! I wonder if it can run on 12GB VRAM. When placing nodes, can they be connected in order? It's hard to follow. Is it possible to keep workflows neat and tidy using groups? Thank you

You should be upfront about how much VRAM this workflow will need. You're not going to be able to run this locally without at least 24GB of VRAM.

Get the same good result without the LoRa. Flux.1-dev dont need any foto-loras. Just prompt what you want to get.

This is impressive work and I'm surprised that there are no comments. Thank you for this powerful UI and integration of Qwen.

Hello, where i can download the workflow?

Was looking for something just like this. Great work!

I want to add this Worklow you shared to Lora, where I trained my own face. How can I use this model finetuned with Lora? How to add it from the menu? Thanks for sharing.

Sorry to say this but from the examples shown they made it only worse. I'd rather prefer someone fixing this annoying banding issue that Flux has. I like Shakker's detail lora though.

hello Everything was ok, but this error is given in the sampler.(object has no attribute Size). How can it be solved?

great, thank you, let me know when you revise the workflow to syn audio and lips. by the way its a very nice workflow. thanks for your effort. Have a nice day SIR !!!

Thanks

I'm seeing so many of these videos for Supir but not a single person has shared the workflows. I am sick and tired of just going through an explanation for this, can someone share the damn workflow for this

Great work, bro. Can it work on my rough sketches?

I try and it freeze at line 136. about 8hrs. 3060/12gb step 1560 8%|▊ | 1/13 [00:00<00:06, 1.91it/s] 15%|█▌ | 2/13 [00:00<00:01, 7.57it/s] 31%|███ | 4/13 [00:00<00:00, 11.23it/s] 46%|████▌ | 6/13 [00:01<00:00, 11.51it/s] 62%|██████▏ | 8/13 [00:01<00:00, 11.49it/s] 77%|███████▋ | 10/13 [00:01<00:00, 11.56it/s] 92%|█████████▏| 12/13 [00:01<00:00, 11.51it/s] 100%|██████████| 13/13 [00:01<00:00, 8.05it/s] [2024-09-15 21:11:14] [INFO] 2024-09-15 21:11:14 INFO move vae and unet to cpu flux_train_network.py:208 [2024-09-15 21:11:14] [INFO] to save memory [2024-09-15 21:11:15] [INFO] 2024-09-15 21:11:15 INFO move text encoders to flux_train_network.py:216 [2024-09-15 21:11:15] [INFO] gpu [2024-09-15 21:11:27] [INFO] 2024-09-15 21:11:27 INFO [Dataset 0] train_util.py:2349 [2024-09-15 21:11:27] [INFO] INFO caching Text Encoder outputs train_util.py:1111 [2024-09-15 21:11:27] [INFO] with caching strategy. [2024-09-15 21:11:27] [INFO] INFO checking cache validity... train_util.py:1117 [2024-09-15 21:11:27] [INFO] 0%| | 0/13 [00:00<?, ?it/s] 100%|██████████| 13/13 [00:00<?, ?it/s] [2024-09-15 21:11:27] [INFO] INFO caching Text Encoder outputs... train_util.py:1143 any idea please

can you check Stable Video 4D?

Are you considering doing a workflow on restoring and coloring old damaged photos?

thanks but honestly useless model, allocation error on 3080 10gb even with 2b and 16 frames , so its better to use ddm tiled workflow example from kijai

it's work, thanks again.

Hello 👋 Can I Use Story Diffusion With Openpose to create consistent animation.. Plz Reply

do you think it is worth to try to use it with a 3060ti 8GB VRAM ?

how can we have a picture with both characters ?

What about image to video?

text to Vedio....? 🤣😂

Nice bro

Hey Buddy Can I use controlnet open pose with story diffusion

Can you tell me the total size of the FluxGym folder after downloading all the models and necessary files? My drive already has low capacity with a lot of various models and LoRAs installed. I just want to make sure I have enough storage left for FluxGym.

FluxGym ui sdxl model traing work?

You added the character at 3:41, then it feels like the video just ends abruptly and we don’t get to see the result of adding that new character

The voiceover in your video says, “The best part is you don’t have to use complicated terminal command”, then it proceeds to show nothing but terminal commands for the first seven minutes of the video. If the viewer has no coding experience, this might be complicated for them

Thank you for the great video~!