∿ Music hackers. Algoraves. Inventing, playing. Neural synthesis. 24/7 ai death metal. Stable Audio team. Open models. Mischief @harmonai_org @artblocks_io
Boston
Joined October 2017
- Tweets8.8K
- Following9.2K
- Followers11K
- Likes15.1K
Pinned Tweet
🥳 Announcing Stable Audio 3 🍕
🏆 fastest music models ever
💻 runs on MacBookPro M-series
🧪 break it plz
🧠 LoRA finetune in < 1h
📷 Sm = faster, Medium = qualityer
⚡ 59x realtime on M5 Pro
One-liner fast install:
curl -LsSf dadabots.com/_/sa3-mac | bash
Meet Stable Audio 3.0, the open-weight model family built for artistic experimentation.
This is our open invitation to experiment with generative audio. We believe the best innovations are still waiting to be built.
The 4-1-1 on 3.0:
📣 You own your outputs, and can distribute and commercialize them under the Stability AI Community License (up to $1 million in revenue).
🎵 New and improved capabilities include variable-length generation up to six minutes, and full song composition on portable devices, no GPU required.
✅ Trained on a fully licensed dataset.
🎨 You can customize the models on your own library with support for LoRa training, which we’ve documented for the first time.
More on the models 👇
dadabots retweeted
What makes a good representation for diffusion models?
Surprisingly, more information isn't always better.
In our new work, Purify Before You Align, we show that removing locally predictable information can substantially improve diffusion alignment.
🎥 A quick overview 👇
🧵 1/n
dadabots retweeted
Fresh take on diffusion distillation: the student learns mean flow by matching the teacher with the sum of a straight-line prediction and a curved zero-integral path at random sub-intervals. For sampling, just throw away the curved part.
Not that simple IMO😁 but very clever!
Excited to introduce Simplified Matching Distillation (SMD) for few-step diffusion🧵
No auxiliary score model. No adversarial training. No generator modification.
You will get:
• Photorealistic appearance
• High-quality dynamics
• Memory-efficient training
• Mode covering with diverse outputs
SMD works well across images, video, joint audio–video, and 3D gen, validated across model scales up to 33B parameters.
Besides, SMD supports:
- Multi-teacher capability merging (like OPD in LLM)
- Cross-model capability transfer
- Bidirectional teacher to causal distillation
Full release. Paper, code, model and data.
Project Page: zjwfufu.github.io/SMD-Page/
Paper: arxiv.org/abs/2610.18472
Code: github.com/zjwfufu/SMD
Model Card & Data: huggingface.co/zjwfufu/SMD
Demo Video 👇
dadabots retweeted
We released the code and model weights here: github.com/ayushjain1144/tra…. Have fun! 🙂
1/ We introduce TrackEverything: a 3D point tracker that tracks all points across all frames of long videos (1000+ frames).
Our key idea is to tie computation to unique 3D scene content rather than redundant 2D pixels in a video.
trackeverything.github.io
dadabots retweeted
Speech can survive being flipped backward every twenty milliseconds.
The same speech in backward pieces of 200 then 100 then 50 then 20 milliseconds.
The original comes last.
When do the words become clear?
A listening demo from @OrukLabs.
underfit in action
things i'm working on today:
- making my song
- trying to fine tune @dadabots' underfit, using claude ultracode to help me get a clean songaday dataset (i've tried this about half a dozen times in the last 2 years on various open source repos to no avail - it all comes out gobbledygook and not in an interesting way)
- making an autoclipper a la opus for the new podcast me and @bigvibessss are starting
- recording with @bigvibessss later
- working on the 3rd level for p(doom), thinking about issuing a challenge for people to be able to earn tokens [2PLF955KTb7CDpHbA5VZH8CgunkvWqfic9wufwJFpump]
- working on a plugin that makes it so i can drop my vocals in and it converts them to speech synthesis computer singing voice (a long held dream of mine)
- integrating some more enemies into wamp to take advantage of our new "boss" feature
- working through a huge checklist of things to add/fix in wamp
dadabots retweeted
Can diffusion models generate high-quality images without CFG?
Yes! We introduce ✨Correction Training for Diffusion Models✨, which significantly improves sample quality without relying on classifier-free guidance.
Project page: sunovivid.github.io/correcti…
🧵 (1/n)
dadabots retweeted
Open weights = open playground 🎵
Huge thanks to @StabilityAI @stableaudio @dadabots and the rest of the open audio crews and more for dropping the weights.
Currently fine-tuning on some beats and having way too much fun with it.
Open source for the win. Always.
(Electronic music) Ok so every album/EP/mix of yours is now just a LoRA dataset. The point of making albums now is just to have the LoRAs. And you mix them live, creating new variations of your music. Every show surprises you. Every live show is now super different and rare. Yet within the variance of your world. Familiar but bizarro.
Your capability for spontaneous improv is expanded. The audience can shout something never before uttered, or shove it your face on their phones, and you can ride the edge of chaos and let it influence the whole set.
Maybe live you try to make your sets different by chopping your tracks on CDJs or remixing stems in ableton. Cool but that's not different the way this is different.
Liminal separatrix spaces exist where it's unobvious where one track (or one artist) begins and another ends, bc the statistical distributions are intertwined. Even with mashups, it's obvious stems are distinct. Here the lines blur. Here you dissolve into the hivemind and reconstitute.
Douglas Adams called probability the fifth dimension. What if you were equipped with the thrusters that propel you at will.
Seems obvious.
This site has a bunch of cool Earthquake Musification Instruments --wait i love this sos.allshookup.org/instrumen…
A lot of ai music i encounter feels like it’s trying to pull me into its illusion.. In fact, the marketing narrative of the services feels that way too..
Where’s the ai music that feels illuminating? breaks free? shows us something real other music could not?
If you’re producing music with ai & machine learning —> lean all the way into it, expose the wires, make it obvious, show us the ugly parts, show us it-being-itself.
Make it glitch. Don’t just generate “glitch music”, or what it thinks is a glitch— show it Actually Glitching. Show it fucking up while trying not to fuck up. Push your process till it rips apart at the seams.
Don’t rear view drive. Take us somewhere new, drive like a crazy person. Throw away the desire to become the idols of the past. Kill the rock star. The DJ. The singer-songwriter.
This is hard. The controls exposed to you want you to rear view drive. It’s literally maybe the only thing it lets you do. It takes some genius to break it.
Do the hard thing. Start with beginner’s mind. Trust your instincts. Reset.
