Joined March 2022
- Tweets 101
- Following 122
- Followers 562
- Likes 488
36 Photos and videos
Zippy retweeted
Mar 23
Uni-1 is here! A new kind of model that thinks and generates pixels simultaneously. Less artificial. More intelligent.
457
776
4,830
6,499,167
20 Aug 2024
Made an fp8 implementation of Flux which gets ~3.5 it/s 1024x1024 on 4090, >2x faster than other methods. github.com/aredden/flux-fp8-β¦ #flux1
3
17
1,487
2 Dec 2023
Finally figured out how to speed up my #sdxlturbo
frontend! It's so fast that the only way to show the actual speed is to delete the prompt, since I can't type fast enough π .. built with next.js frontend & tensorrt backend.
15
28
217
43,754
2 Dec 2023
For anyone who may think that this is a fast-forwarded video, or using image caching, I can guarantee it is not- the images are all base64 and it's using a websocket, the terminal below is the output from the tensorrt rest-api server in real time as I ctrl y / ctrl z
3
23
1,453
30 Nov 2023
Now we're real time! π - with SDXL turbo & tensorrt optimized further.
6
2
20
1,528
1 Dec 2023
Now I'm at 38 fps, from switching to taesd- but have no way to view the images that fast π₯Ή
2
6
314
30 Nov 2023
Optimized further meaning I'm getting 17 fps for full clip encode -> unet pass -> vae decode -> serialize to jpeg bytes.
5
269
29 Nov 2023
Made a neat GUI in react for fun & hooked it up to an optimized sdxl-turbo tensorrt api backend I built for image autocomplete! It generates so fast that my browser cant keep up π₯² - so fun! #sdxlturbo #sdxl #AIart #stablediffusion
3
5
17
2,267
9 Jul 2023
SDXL unet finetuning is going pretty well! FIgured out how to train without deepspeed on 4090s π
#SDXL09 #diffusers #AIArtwork
2
10
907
5 Apr 2023
Blockadelabs built a neat thing! π₯³ (which I built the ML backend for) π³
5 Apr 2023
REEEMIXXXXX! π’π’π’
Infinite 360Β° worlds just got easier to fine tune. Transform anything - including visual style, time of day, & weather while keeping the structure intact!
LIVE NOW: skybox.blockadelabs.com
#skyboxlab #texttoskybox #aiart #gamedev #aitools
4
3
30
4,229
We release a new ViT-G/14 CLIP model with OpenCLIP which achieves 80.1% zero-shot accuracy on ImageNet and 74.9% zero-shot image retrieval (Recall@5) on MS COCO. As of January 2023, this is the best open source CLIP model.
laion.ai/blog/giant-openclipβ¦
huggingface.co/laion/CLIP-Viβ¦
16
160
805
298,818
15 Jan 2023
Will probably be hard to appreciate via twitter, but- here is a true seamless 15616 by 1024 image 100% generated by stable diffusion π€―
12
8
67
11,788
15 Jan 2023
Here is a closeup of the right end, so you can get a sense of detail
5
496
1 Jan 2023
12
1,067