Radamés Ajna PRO
AI & ML interests
Articles
Organizations
radames's activity
Congrats and welcome to the team!
posted on X https://x.com/radamar/status/1798817359224025525 @Jofthomas do you have an X account? Maybe you could push it on LI? thanks
thanks to @Jofthomas now you can login with HF and talk to agents or other humans
Amazing team! Looking forward to plug my own tools!
Amazing work @ehristoforu . Do you have plans to open-source your training recipes?
Here's a demo for you: radames/MistoLine-ControlNet-demo
Super resolution version: radames/Enhance-This-HiDiffusion-SDXL
from controlnet_aux import AnylineDetector
anyline = AnylineDetector.from_pretrained(
"TheMistoAI/MistoLine", filename="MTEED.pth", subfolder="Anyline"
).to("cuda")
source = Image.open("source.png")
result = anyline(source, detect_resolution=1280)
yes that's the idea, you Sign in with HF and it syncs the db with personal dataset , in case you reboot it I'll look back your state. 👀 into it
You can learn more about it here https://huggingface.co./blog/radames/hugging-face-google-visual-blocks
Source-code for the custom nodes:
https://github.com/huggingface/visual-blocks-custom-components
Yes that's a great idea, I'm chatting with folks from Convex, and check if the sqlite db is the only file I need to backup, then I'll set a scheduler to push it to a personal dataset. Folks them would be able to pause and restart from that state!
Thanks! It was a fun challenge to put it all together in a single container. I'm excited to try more Convexdb as a vector db and backend.
Demo: radames/ai-town
Instructions: https://github.com/radames/ai-town-huggingface
Just saying that I really like the ability to quickly test your model against the monster ones! It's amazing how well it performs against Claude. 🤯
Amazing!! Shall we make a VB node for this?
Demo: radames/Enhance-This-HiDiffusion-SDXL
Older version based on DemoFusion radames/Enhance-This-DemoFusion-SDXL
New Controlnet SDXL Controls Every Line TheMistoAI/MistoLine
HiDiffusion is compatible with diffusers and support many SD models - https://github.com/megvii-research/HiDiffusion
Thanks for the summary,
@Sentdex
!
For the curious, there are some examples on their GitHub repo https://github.com/KindXiaoming/pykan
Hi @renyuxi , thanks for sharing this update! 8 steps with CFG and negative prompts is amazing!
Basic snippet
# pip install gradio_rerun gradio
import gradio as gr
from gradio_rerun import Rerun
gr.Interface(
inputs=gr.File(file_count="multiple", type="filepath"),
outputs=Rerun(height=900),
fn=lambda file_path: file_path,
).launch()
More details here radames/gradio_rerun
Source https://github.com/radames/gradio-rerun-viewer
Follow Rerun here https://huggingface.co./rerun
amazing work @oliveryanzuolu 👏 Do you have plans to release the training distillation code?
Here a few demos:
Official:
Hyper-SDXL-1Step-T2I ByteDance/Hyper-SDXL-1Step-T2I
Hyper-SD15-Scribble ByteDance/Hyper-SD15-Scribble
Unofficial Demos: InstantStyle + Hyper SD1.5 (not great but super fast) radames/InstantStyle-Hyper-SD
InstantStyle + Hyper SDXL radames/InstantStyle-Hyper-SDXL
In a big related update, as of today, Diffusers@main supports InstantStyle. I'm looking forward to playing with it!
https://github.com/huggingface/diffusers/pull/7668
radames/InstantStyle-SDXL-Lightning
ByteDance/SDXL-Lightning
Very interesting, @andrewrreed , and completely unaware of this feature! Do you know of any other strategies for grounded generation in models like LLaMA or Mistral?
pip install gradio_huggingfacehub_search
You can see it in action here. arcee-ai/mergekit-config-generator
And learn how to use it here radames/gradio_huggingfacehub_search
radames/Candle-Moondream-2
ps: I have a collection of all Candle WASM demos here radames/candle-wasm-examples-650898dee13ff96230ce3e1f
nice!! can you set the jpeg quality as well?
thanks @chansung this is so helpful! btw you could launch a quantize version here https://huggingface.co./spaces/chansung/gradio_together_tgi/blob/main/entrypoint.sh.template#L11 and even try running this on CPU.
@Wauplin is doing impressive work here. 👏
It's very interesting how ControlNet Canny quality is comparable, but in a single step. Looking forward to when they release the code: https://github.com/GaParmar/img2img-turbo/issues/1
I've been keeping a list of fast diffusion model pipelines together with this real-time websocket app. Have a look if you want to test it locally, or check out the demo here on Spaces.
radames/real-time-pix2pix-turbo
Github app:
https://github.com/radames/Real-Time-Latent-Consistency-Model/
You can also check the authors img2img sketch model here
gparmar/img2img-turbo-sketch
Refs:
One-Step Image Translation with Text-to-Image Models (2403.12036)
cc @gparmar @junyanz
hi @visheratin , do you have any guides on how to train similar model? Phi-2 + SigLIP vision encoder?
I know it's possible to run real-time whisper on a rapberrypi with whisper.cpp @ggerganov
Are you thinking of running it on a device or in the cloud?
hello 👋