Consumption acc/acc
@c0nsumption_
Followers
4,620
Following
175
Media
1,857
Statuses
4,975
CTO at @PlayhouseAI Exodus 3:14 #RedTeam 🟥 悟り 💀🔫 c0nsume. 悟る
Joined January 2022
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
President
• 4564414 Tweets
Kamala
• 4105374 Tweets
Joe Biden
• 2310328 Tweets
Democrat
• 663462 Tweets
The Loyal Pin Trailer
• 396408 Tweets
#HouseOfTheDragon
• 258037 Tweets
Rhaenyra
• 150009 Tweets
Manchin
• 142225 Tweets
Miguel
• 137754 Tweets
Seasmoke
• 78376 Tweets
#GranHermanoCHV
• 75389 Tweets
東海道新幹線
• 70298 Tweets
#HOTD
• 69849 Tweets
Daemon
• 67570 Tweets
Leah
• 60063 Tweets
Jana
• 56272 Tweets
Mysaria
• 49007 Tweets
Alicent
• 44933 Tweets
Punk
• 43211 Tweets
Kordell
• 43015 Tweets
Heihachi
• 38235 Tweets
#LaCasaDeLosFamososMx
• 35750 Tweets
Addam
• 27331 Tweets
ポポロ休刊
• 24922 Tweets
Kendall
• 23404 Tweets
Aemond
• 22685 Tweets
भगवान शिव
• 21399 Tweets
#ピザBENTO
• 16036 Tweets
Yuhui
• 13011 Tweets
北陸新幹線
• 12179 Tweets
電波人間
• 12113 Tweets
Harrenhal
• 12017 Tweets
Pinned Tweet
✋🏽💀 'And there will even be audio.'
I vividly remember living in a car.
I vividly remember no one.
And I vividly remember:
"🖕🏽 ONLY UP FROM HERE. "
0
0
2
SDXL Turbo to SD15 as refiner
🤔 idk what’s more impressive:
• how much SDXL Turbo fine tunes drastically improve generation quality
• SD15 can refine a 1024 x 1024 image
25
90
710
＼（〇_ｏ）／Made a video tutorial for animatediff-cli-prompt-travel. It get's you up and running and I included a prompt.json file to make things easier for beginners.
#AnimateDiff
#prompttravel
#AIArtCommuity
Link here:
37
87
643
(￣y▽,￣)╭ I am happy to annouce:
The start of my Github Repo of Workflows is made. Tis' only the beginning comrades.
Currently Available:
- SDXL to SD1.5 refinement
- First to Last Frame Interpolation (AnimateDiff)
More is on the way 😈!
Link:
26
90
608
animatediff-cli-prompt-travel with IPadapter (allows for image prompts to mix with text prompt). Increases the consistency of generation. Tutorial is on the way. Now experimenting with mixing in LoRA's with IPAdapters.
#Animatediff
#生成AI
#AI
11
53
453
"Did it work?"
Yes. Yes it did.
⚠THIS IS NOT Video to Video⚠
- this is the raw output of a MotionLora
- it was trained on the footage on the left.
Now its time to really test.
#MotionLora
#MotionDirector
#AnimateDiff
#ComfyUI
21
66
378
ヾ(￣▽￣) 🤚🏽 Going to be releasing my ComfyUI workflow and tutorial tomorrow
明日、私のComfyUIのワークフローとチュートリアルをリリース予定です。
Voy a lanzar mi flujo de trabajo y tutorial de ComfyUI mañana.
#AnimateDiff
#ComfyUI
#AI
アート
10
31
364
（づ￣3￣）づ And now: ControlNet Powered Video To Video!!!
There's a lot more on the way. Have to work now but stay tuned for IPAdapters and Keyframes. Releasing video maybe tonight!
#AIArtwork
#VFX
#AnimateDiff
#AIart
#AI
アー
14
51
356
(￣y▽￣)╭ And here we have: Unlimited Animation Lengths in ComfyUI AnimateDiff...
And the tutorial is only 2 minutes long 😎
Coming next? 3 image interpolation w/ prompt travel and IPAdapters.
#aiart
#AnimateDiff
#AI
アート
15
44
339
So we can do 3D Gaussian Splatting in ComfyUI.
But WHY?:
- Gaussian Splatting can be used for shot reframing
- Creating meshes (although still needs work)
- Rapid product conceptualization via 3D printing
The research is surely on its way.
#3DGaussian
#GuassianSplat
#ComfyUI
6
53
301
[]~(￣▽￣)~* Behold! I now bring the Install & Setup video for ComfyUI w/ Animatediff Prompt Travel:
Video on IPAdapter's (my workflow) coming next 😎
#AIart
#Animatediff
#AI
アート
9
45
302
LCM + MotionDirector LoRAs is:
- fast
- powerful
- a lot of room for creativity
- still a lot to be tested and discovered
I'll be re-filming my scenes this week using my new research. Tired.
Onward 🏇🏽
#lcm
#MotionDirector
#MotionLoRAs
20
31
283
(￣y▽,￣)╭ I cannot emphasize enough: Stable Video CAN MOST DEFINITELY BE PROMPTED‼‼‼
11
24
237
（*゜ー゜*） Sound On Please!
ComfyUI workflow for turning SDXL Turbo outputs into a video! Pass it folder of SDXL Turbo frames or any folder of frames.
This piece is called: 'God of Time'
Audio by me as well.
Workflow here:
15
33
234
And thanks to ComfyUI we can merge SDXL Turbo with existing SDXL model. This increases generation quality drastically.
@StabilityAI
amazing 🧍🏽♂️
#SDXLTurbo
12
27
227
Huge shout out to
@Lykon4072
The creator of DreamShaperXL who already dropped a SDXL Lightning Finetune.
4 steps looking amazing.
I guess SDXL is fast now 🤷🏽♂️
@c0nsumption_
try that in pair with Turbo (both at lower ratios). This one was roughly 1s to generate and 1s to upscale to 1224x1632 at no quality loss compared to the base model.
4
1
14
5
26
214
BEHOLD o(￣▽￣)ｄ
The next video tutorial: IPAdapter (Image Prompts), LoRA, and Embeddings. Prompts file supplied and links in description. Let's keep the
#Animatediff
train rolling.
9
13
202
Can SDXL Turbo to SD15 be a refining tool that can be prompted with color management control?
Yes.
Swap VAEs on the fly. Tested:
Berry’s Mix
Orange Mix
kl-f8-anime
kl-f8-anime2
vae-ft-mse-84000
Prompted for oil paint on canvas.
SD15 or $39 a month? You know who you are. 🤫
8
21
207
@DiffusionPics
For anyone interested: This is my work. I’ve made a tutorial if you are interested on getting set up. Currently making the next vid explaining IPAdapters.
Also check the original Reddit post for more info:
＼（〇_ｏ）／Made a video tutorial for animatediff-cli-prompt-travel. It get's you up and running and I included a prompt.json file to make things easier for beginners.
#AnimateDiff
#prompttravel
#AIArtCommuity
Link here:
37
87
643
5
13
197
(￣o￣) . >"Sky Walker" not the Star Wars kind.
So much to learn it's unreal.
Getting the hang of posting on 4 platforms.
Quality has increased.
Now it's all about increasing frequency.
Thanks for your patience.
#LCM
#AnimateDiff
#aivideo
12
12
190
And this is why I love ComfyUI.
Not even full day. We just got the Gradio demo working haha.
👇🏽LINK IN THREAD 👇🏽
#TripoSR
#3dmodeling
#ComfyUI
9
21
180
Automatic Asset Generation.
Pipeline:
SDXL Layered Diffusion -> TripoSR
- prompt desired asset
- cycle for a decent output
- clean up in 3D software of your choice
@MrForExample
I'm dropping the workflow in my repo, just cleaning it up
Onwards 🏇🏽
#3Dartist
#3DMesh
4
21
172
Alright, now testing actual SDXL-Lightning LoRA.
8 Step here. This is an actual SDXL Finetune running at 8 steps. 🤯 Quality loss seems minimal considering my history with this prompt.
#SDXL
3
16
166
(づ￣ 3￣)づ Workflow File up:
- prompt-able 3D asset generation
- generation time on 4090 like 15 - 30 sec
(dependent on geometry_extract_resolution)
\
- circumvents saving bug w/ name templating
👇🏽Link in thread👇🏽
#Nerf
#3DModel
#3DPrinting
#gamedev
#animation
9
25
166
AnimateDiff finally hooked up. Looks good but lost some facial coherence. No biggie, fixing now.
Headshot -> Fullbody -> Video Scene.
Almost there. Onward.
#ComfyUI
10
19
160
SVD1.1 Weights Dropped?
Downloading now.
@StabilityAI
Excited to see how it plays into my current system.
#SVD
#ComfyUI
5
20
157
( ﾉ ﾟｰﾟ)ﾉ And NOW!: Cloud GPUs for ComfyUI AnimateDiff Prompt Travel for all my folks without a 4090!
Now we can get on to all the good stuff. You're gonna need VRAM for all the gold coming next 😉
#AnimateDiff
#AIart
#AI
アー
#AIArtwork
#VFX
13
17
150
New Open-sourced background segmentation model.
Results are EXTREMELY good.
I mean that. And this is the worse it’ll ever be so…
Expect EVEN BETTER with time!
#segmentation
#sam
#bgremover
Excited to share we just open-sourced our new background segmentation model 🥳
🚨 Check out our
@gradio
demo
RMBG v1.4 by BRIA excels in separating foreground from background across diverse categories, surpassing current open models 🚀
13
92
391
6
17
149
＼（〇_ｏ）／ "Start you engines!!"
@StabilityAI
Released the models for local use. Gotta love that they exist in the AI space.
6
26
145
Someone on Reddit replaced Stage B of the Cascade pipeline with SDXL and SD1.5.
The results are actually pretty dang good.
👇🏽Post in thread 👇🏽
(They include the workflow)
#StableCascade
#CustomPipeline
#SDXL
5
20
145
ComfyUI 3D Pack Running 🙏🏽
Nodes created by
@MrForExample
Windows 10 install works.
No luck on Ubuntu 22.04
👇🏽LINK BELOW!!👇🏽
#3DModel
#3Dai
#ComfyUI
#3DGS
4
17
142
Should most def check this out everyone.
OP says it works on 12 GB cards.
👇🏽Repo in thread! 👇🏽
#Upscale
#magnificAI
#opensource
12
14
141
Good news, my pull request has been merged into the gradio app for TripoSR repo by
@tripoai
The gradio app should no longer give any issues :)
GET ON IT.
This thing is awesome!!!
Look how quick it is!
#3DGS
#research
#3dmodeling
#NERF
#AIArtCommuity
#gamedev
7
15
138
SDXL-Turbo. Funny thing is it can go even faster. Pretty much prompting in real time. Here I have the seed also randomly generating as well.
#SDXLTurbo
6
14
143
Stable Video 1.1 looks extremely promising.
- less blur
- believable animations
- grounded camera movements
With my video repair tool, these outputs are becoming close to useable. DMs are open.
#StableVideo
#ComfyUI
11
12
137
￣へ￣ Finally.
So much studying.
So much learning.
Music Video is going to be x100 times better now.
AnimateDiff tis the way.
#vibrant
#scifi
#animatediff
#lcm
8
11
126
Walking Drone Shot Motion Lora
Test
#2
Again:
- rotoscope & place in desired scene.
Quality will increase over time as well.
We've just cracked open
#MotionDirector
.
Imagine a month from now?
#MotionLora
#MotionDirector
#AnimateDiff
#aiartcommunity
#animation
8
12
122
Someone already dropped a fine tuned SDXL Turbo model....
(*￣０￣)ノ Introducing: TurboVisionXL by socalguitarist on civitAI
5
14
122
(☞ﾟヮﾟ)☞ And now: Tutorial for ControlNet Keyframe interpolation in animatediff-cli-prompt-travel
AnimateDiff-cli-prompt-travelでのControlNetキーフレーム補間のチュートリアル
Keyframes used attached in comments.
#AnimateDiff
#AI
#AI
アート
5
13
114
(☞ﾟヮﾟ)☞ ComfyUI: Installation & First Image
🎥Check out my new video!
- covers installation and setup of ComfyUI
- installing the ComfyUI manager
- generating your first image
This will be a zero to expert series, ensure to tag along!
FULL LINK:
4
18
121
SVD1.1 Looking dope.
This isn't even in my pipeline yet!
Results look promising.
#SVD
#StableVideo
2
6
113
The Facial Refiner in my workflow.
Thanks to
@javilopen
In trying to replicate MagnificAI, the things I am now able to accomplish in ComfyUI are remarkable.
You have one hell of a product not that I've ever used it. The push you've given to us devs to compete is a blessing👌🏽
6
10
109
1 STEP SDXL-LIGHTNING IN COMFYUI!!!!
Yo. I am so damn hype :D
This is going to be such a powerful tool.
This is the UNet model:
sdxl_lightning_1step_unet_x0.safetensors
I am adding the workflow to my repo!!
#SDXL
#Lightning
4
11
105
AnimateDiff v3 tutorials dropping this week.
Livestreams on YouTube and X.
Get everyone up to speed on:
- v3 parameters and nuances
- LCM work
Also been working on some useful blender tools.
More on that later.
#AnimateDiff
#LCM
8
12
104
You know... interpolating in ComfyUI gives Topaz a run for its money 🤔
#stablediffusionvideo
6
10
102
Stable Diffusion 3
HunyuanDiT v1.1
PixArt
and also: Lumina-T2X
Another Diffusion Transformer (DiT) based architecture:
#airesearch
1
22
100
Motion Director working really good.
Motion Lora is holding the concept nicely.
Which means this week will be a fruitful week for gathering footage:
- that I film so that it's properly sourced
- useable in common film scenarios
#Training
#AnimateDiff
#film
#animationdev
3
6
96
SDXL-Lightning Unet Models also work with SDXL finetunes. Love it.
🤔Shall we check LoRAs?
#SDXL
#UNET
#Lightning
4
10
95
Alright on a
#Nvidia
3090TI:
Batch Size of 4.
Images Generated at roughly 5 seconds.
Thats nuts.
On a 4090 will be even faster.
Will test later.
#AMAZING
#SDXL
#Lightning
10
7
90
"Where's your music video?"
( ˘︹˘ ) I had to detour.
Learned some huge video stuff.
And the results speak for themselves.
Was worth it.
Onward \(￣︶￣*\))
#AnimateDiff
#MotionDirector
#MotionLora
#Finetune
5
8
89
Sorry X! Super drained after Muay Thai training. Going to bed early today (´。＿。｀)
Forgive me. I promise I have a some goodies and updates coming tomorrow.
- SDXL turbo workflows
- Testing the new SD2.1 Turbo model
- as well as filming new tutorials 😉
Good night!!!
6
4
89
The Mixtral of Stable Diffusion?
“SegMoE is a powerful framework for dynamically combining Stable Diffusion Models into a Mixture of Experts within minutes without training.”
🤯Insane.
@NerdyRodent
made a video covering this:
2
17
89
Remote Display (Beta) by
@MetaQuestVR
just dropped. Wirelessly mirror your desktop display to your quest.
Yes, you’ll still be able to use quest apps simultaneously like the web browser on the right!
#SpatialComputing
#RemoteDisplay
9
14
85
＼（〇_ｏ）／ Dalle3 input fed to
#StableVideo
via ComfyUI. Interesting results for sure. Looking forward to more testing. Pooped for today
5
9
84
(づ￣ 3￣)づ New tutorial is up:
Inpainting With Differential Diffusion.
If you plan on keeping up with my recent work:
- Animatediff
- MotionLoras
- 3D Gaussian Splats
Please ensure to follow along.
⚠ LINK IN THREAD 👇🏽⚠
#ComfyUI
#Inpainting
#tutorial
#aiartist
6
6
84
So… an AI Upscaler that can also refine images and be prompted? Looks like Krea AI beat me to it 🤔
Genuinely impressed. Learning the ins and outs of the app. So far I’m liking what I see.
3
8
80
7
17
81
ZeroScope V2 XL running in ComfyUI.
Met the dev, extremely skilled individual.
Gotta diversify the tools.
Learn a lil of everything.
So here's Darth Vader Surfing.
Onward 🏇🏽
#surfer
#modelscope
#zeroscope
#ComfyUI
2
7
81
For anyone interested in getting
@playground_ai
new model running in ComfyUI:
I've added the workflow to my repo.
it can be found here:
#ai
#playgroundai
#newmodel
6
9
77
Playgroundv2.5 Refiner Workflow.
SDXL Lightning Fine Tune Based.
Removed steps from initial pass.
Renders in a similar amount of time.
👇🏾WORKFLOW IN THREAD👇🏾
#PlaygroundAI
#Refiner
#SDXL
#Lightning
5
7
75
(☞ﾟヮﾟ)☞ Behold:
⚠ MOTION DIRECTOR! 🚨
- step by step set up
- evaluate each stage of training
- explain crucial concepts
- follow up video coming today
I know it's what you've all wanted 😈
👇🏽 LINK IN THREAD 👇🏽
#ComfyUI
#Inpainting
#tutorial
#aiartist
#MotionDirector
#AI
2
9
74
2
10
73
Initial Benchmark Print of an output from TripoSR developed by
@tripoai
&
@StabilityAI
3D Printed with a layer height of .20mm on
@Creality3dP
K1Max
Can improve it easily via re-meshing and some modeling work.
…But challenge is to only use AI.
#mesh
#ai
#3DGS
#nerf
#aiart
Good news, my pull request has been merged into the gradio app for TripoSR repo by
@tripoai
The gradio app should no longer give any issues :)
GET ON IT.
This thing is awesome!!!
Look how quick it is!
#3DGS
#research
#3dmodeling
#NERF
#AIArtCommuity
#gamedev
7
15
138
3
8
75
"cloud of atoms"
The idea that we are a "cloud of atoms" with new ones entering and old ones leaving: the dynamic and ever-changing nature of matter at the atomic level.
Powered by AnimateDiff.
#fire
#aianimation
#atoms
#Motion
2
8
71
TripoSR:
GPU Usage is directly proportionate to:
- ChunkSize (Default: 8192)
- Geometry Extract Resolution (Default: 256)
Highest useable value for Geometry Extract Resolution was 390.
Posting associated error in both repos.
#3DPrinting
#mesh
#3DGS
#nerf
#ComfyUI
2
10
70
ψ(._. )> ComfyUI Differential Diffusion Inpainting.
Quick and dirty face change (swap? 🤔).
Nothing but the nodes on screen.
Dropping in the repo ｓｏｏｎ 💀
#ComfyUI
2
7
71
Motion Director is 🔥
- it only takes 15 min+ to train a motion
- drastically more control over the scene
- this isn't even following best training practices and the results are pretty good..
Using weekend to test a bunch of hypotheses.
#Animatediff
#text2scene
9
3
69
3
9
69
5
9
67
Stable Cascade
- this thing is amazing
- easily a huge Graphic Design tool
- is one hell of an image generator
You can tell a story with thing.
#Stable
#Cascade
#storytelling
4
4
65
(*￣3￣)╭ This VIDEO REFINER is *chef's kiss* 👨🏽🍳
🎥Stable Video 1.1 proving to be a powerful video signal generation tool when paired with my refiner.
Onward 🧍♂️
#StableVideo
7
5
62
@aisolopreneur
This is my work everyone. This keeps on happening and it’s getting annoying. Please tag original creators.
＼（〇_ｏ）／Made a video tutorial for animatediff-cli-prompt-travel. It get's you up and running and I included a prompt.json file to make things easier for beginners.
#AnimateDiff
#prompttravel
#AIArtCommuity
Link here:
37
87
643
5
4
61
"Gaze into the eyes of a Samurai"
Do nothing which is of no use.
- Musashi Miyamoto
Onward 🏇🏽
#animatediff
#storytelling
#aivideo
#Samurai
3
9
64
So much innovation yet to come.
🔥 Long Input length Increase the maximum input length of CLIP from 77 to 248.
1
12
61
So
#Sora
is able to interpolate between two videos.
BUT LOOK AT IT:
It's as if it's interpolating the pixels AND THE CONCEPTS. That's mind blowing, the ease in and the consistency.
The Art everyone will be dropping is going to be addicting.
#AIArtCommuity
6
7
57
The dev of ComfyUI-3D-Pack just integrated TripoSR
@MrForExample
is on a roll and I genuinely believe this will the definitive ComfyUI 3D suite.
Currently developing some workflows to increase generation quality in this system!
#nerf
#3DGS
#3dmodeling
#3dart
#ComfyUI
#Comfy3D
@VastAIResearch
@c0nsumption_
Comfy3D Update:
- Integrated TripoSR
- Add pre-build wheel for torchmcubes (windows, python 3.11, cuda 12.1)
Is an image to NeRF to Mesh pipeline, there should be better method that marching cube though, gonna keep research!🍻
8
40
270
2
10
58
SVD last frame used to generate SVD output. Trying to debug discoloration issue. Have tested passing output through AnimateDiff and also passing through KJNode Colormatch. They are posted in this thread
#StableVideo
1
5
57
@DMC_Ryan
I was crappin myself when they rumored $2500. That $3500 hit like a slug to the chest.. like an arrow to the knee.. like a “We have to talk” on the morning of your birthday..
$3500… that’s two 4090s on sale.
0
1
56
Stable Video Diffusion can be controlled with prompts 💯
It a question of how much.
#StableVideo
#AIart
1
10
56
（*゜ー゜*）Forgive my delay! Here we have LoRAs and Embeddings:
ControlNet 3 init image, IPAdapter, Refiner, And Frame Interpolation are on the way. I had to account for users with VRAM limitations! Custom node group for this!
#AnimateDiff
#AIart
#AI
アー
4
8
51
（*゜ー゜*）My take on Stable Video?:
The more I use it the more it just blows me away🤯
6
5
53
After much discussion with
@PurzBeats
I will be switching up my video style to live streams and showcases of my work rather than edited and cut tutorials. They are very draining to produce 🧍🏽♂️
I make music, art and videos, so I will begin sharing a lot of me
#AnimateDiff
#AIart
5
2
52
"To meet an Ancestor"
Communicating with the dead has been both a philosophical puzzle and a cultural practice causing much argument and debate around its possibility
Could AI connect us with the dead?
Or some other spirit realm?🤔
👇🏽Thoughts?👇🏽
#AI
#spirituality
#FutureTech
7
4
52
@StabilityAI
“We’d like to announce that we still haven’t released the weights and are just trolling at this point. Oh, and decentralize ai 🤣🫱🏽🫲🏾”
- StabilityAI at this point.
0
0
50
But can it Motion Lora? Test
#2
I'm a Full Stack Software Engineer with a passion for this.
Huge shout out to
@peteromallet
, creator of banodoco.
Again, we bout' this out here.
No talk. Action.
#onwardsandupwards
#drone
#Pilot
#AnimateDiff
#MotionDirector
4
1
48
And here I have that same prompt as before.
But this time an SDXL LoRA is in the chain.
So:
SDXL > SDXL LoRA > SDXL-Lightning LoRA
The results are 🤯 for 8 steps.
Not hyping it up:
I AM GENUINELY HYPE.
#SDXL
#AIArtCommuity
4
7
47
Getting there.
Dolly Zoom Motion Lora.
I'd Say at this point both were a success.
Just about fine tuning parameters.
Onward 🏇🏽
#AnimateDiff
#AICameraGuy
#FashionTrends
#styleoftheday
3
2
47
Link:
Sharing resources for Auto Captioning to my Github this week. Will include:
- BLIP 3
- CogVLM
- Possibly InternVL
Also: Some LumaLabs chaos 💀🥰🥒
#lumadreammachine
#VLM
2
8
47
2
4
44