Amuse is trying it's hardest to censor perfectly fine frames of a woman :D
@OcihEvE29 күн бұрын
Just found Amuse a few days ago, purely by accident. I'm trying to figure out if it supports Lora and I just haven't figured out how or if it doesn't.
@HexOxide29 күн бұрын
Amuse uses ONNX for its models, which are pre-compiled and optimized for inference. Since LORAs require dynamic modification of model weights during runtime, they are not compatible with ONNX models and therefore unfortunately cannot be used with Amuse at this time.
@OcihEvE29 күн бұрын
@@HexOxide It's a bit strange though. I've pulled checkpoints off Civit and drag and dropped them in to the Amuse folder and they worked. I'm thinking if they add a noodle (I think that's the term being used) in their back end for Lora, it would work.
@HexOxide29 күн бұрын
@@OcihEvE If the checkpoints worked in Amuse then they must still have been converted or compiled to the ONNX format.
@iulian80xmusic2 күн бұрын
The download from the website doesn't work can someone link me the app ?
@HexOxide2 күн бұрын
@@iulian80xmusic Weird, should be working? www.amuse-ai.com/downloads/Amuse_v2.3.2.exe
@Gluosnis9Ай бұрын
48 GB vram? What gpu is that?
@HexOxideАй бұрын
AMD Radeon Pro W7900
@D0nKeYАй бұрын
Didnt really work did it. It just created a whole new image. The whole image is changed not just the masked part.
@gamesanimesaleatorio2 ай бұрын
still crashing on my Rx 6600... when they gonna release a version to series 6000 AMD video card?...
@hubtv17402 ай бұрын
grazie
@cocosloan37482 ай бұрын
Keep them videos comming @ HexOxide ❤
@cocosloan37482 ай бұрын
What GPU you are using ?
@cocosloan37482 ай бұрын
How to use video to video ? I saw your vid using Pixelate filter but I want to use something like Cartonize filter - so same girl dancing but cartoon . Good luck with your channel - audio is a bit low but fine ...👍👍
@CGFUN8292 ай бұрын
the midle tab on top that says video generation is the one u looking for
@dp60122 ай бұрын
I'm trying to recreate the same images as in EZ mode of text image but now that I'm in advanced mode I get totally different versions. What's the trick here? I've experimented with some settings but I don't have a result like in EY mode of text image. By the way, good video.
@CGFUN8292 ай бұрын
good stuff man , now i know!
@alberto20542 ай бұрын
Can you make any tutorial about Video Generation?
@Slapdoge3 ай бұрын
Doesn't work for me, it finishes upscaling but there is no image output
@INFO-DIY3 ай бұрын
MY PC => onnxruntime failed to initialize amuse. HELP ME
@paulehigie98793 ай бұрын
what gpu are you using
@HexOxide3 ай бұрын
AMD Pro W7900
@paulehigie98793 ай бұрын
@@HexOxide thought you were using dual 7900 xtx. Wish i could get both of mine to work as one
@kalidesu3 ай бұрын
You voice is barely audible.
@aminshobeiri89013 ай бұрын
Where is the negative prompt box?
@HexOxide3 ай бұрын
Not all models use a negative prompt.
@roki9774 ай бұрын
is that w7900 pro? Its super fast for Amuse and Flux.1.. I generating on 7800xt and 16gb is not enough for Flux.1. It works but it take tons of ram and slows down to like 2,3 mins for picture. Other models work quite fast but they are not so good like Flux.1 so its worth waiting for me. Amuse is great i love it.
@HexOxide4 ай бұрын
Yeah, I've had the luxury to be recording these videos using a W7900 Pro, my own PC has an old 2080TI that is slowly dying and can't run Flux so I feel your pain. One trick you can use if you don't like waiting the 2-3 minutes between each generation, is once you've zeroed in on the prompt and other settings you want to use, to set up the automation to batch 100 or 1000 images to leave running over nigh and then in the morning pick out all the best results and then discard the rest.
@mirza-p8j4 ай бұрын
Why did the entire image change? Out of the mask
@mirza-p8j4 ай бұрын
When will appear openpose feature ectractor?
@zalztorus50954 ай бұрын
Hey do you know if there is a way to import other models not listed there, i.e. from huggingface or civitai.? I tried downloading some but they have various file formats like .ckpt and .safetensors that dont seem to be compatible with Amuse as they wont Import.
@TheBasketballOdysseyАй бұрын
Been researching and found this, "Amuse 2.0 primarily utilizes a proprietary model format, closely tied to AMD's XDNA technology, which is not publicly accessible but is specifically optimized for AMD hardware, particularly the Ryzen AI series processors and Radeon RX 7000 series graphics cards" So we are out of luck on using our own models for the foreseeable future.
@zalztorus5095Ай бұрын
@@TheBasketballOdyssey thanks, I tried a converter after posting this but never had any luck, that seems to be why
@758185luan4 ай бұрын
My picture is blur 😢
@sdwold4 ай бұрын
thats the NSFW filter. cant turn it off
@slim420-e8v3 ай бұрын
Download the filter remover.
@sdwold3 ай бұрын
@@slim420-e8v what filter remover???
@slim420-e8v3 ай бұрын
@@sdwold kzbin.info/www/bejne/aGXMaWuLdqmGfrc
@GilmarcioZuilaneАй бұрын
@@slim420-e8v ?????
@Mrrabbit474 ай бұрын
how to generate bobs and vajayjay
@758185luan4 ай бұрын
my photo is blurry when created can you help me
@kyzer973204 ай бұрын
probably NSFW
@758185luan4 ай бұрын
@@kyzer97320 thank man
@sandwichninja3 ай бұрын
AMUSE AI has extremely restrictive censorship. Censored images get blurred. I makes the product basically useless since so many of the models are trained on nude images. So they will constantly generate lewds that are instantly blurred by AMUSE even when you're not trying to create lewds. It's garbage.
@RichoPOV5 ай бұрын
Hell yes Hex, another banger!
@Mrrabbit475 ай бұрын
that was legitness
@tomasdelatoba7572 Жыл бұрын
Wtf, never would have seen this coming but ai Cartman does spit it