→ @dadabots ← Creative AI & Death Metal #DSP #MIR #NLP #ML #NeuralNetworks | @Harmonai_org
Boston
Joined July 2009
- Tweets1.4K
- Following594
- Followers855
- Likes1.1K
Pinned Tweet
Our paper "Generating Black Metal and Math Rock: Beyond Bach, Beethoven, and Beatles" got accepted by NIPS 2017 Workshop for Machine Learning for Creativity and Design! @zzuk_at_aol #dadabots
CJ Carr retweeted
WizardCoder-15B is crushing it‼️🔥
Amazing progress in open-source code LLMs recently. I believe open-source may have progressed beyond closed-source models in this area.
github.com/nlpxucan/WizardLM
🔥🔥🔥
Introduce the newest WizardCoder model !
1. Surpass Claude-Plus (+6.8), Bard (+15.3) and InstructCodeT5+ (+22.3) on HumanEval Benchmarks
2. Paper is coming, with brand-new Evol+ methods for code LLMs
Github:
github.com/nlpxucan/WizardLM…
HF Weight:
huggingface.co/WizardLM/Wiza…
CJ Carr retweeted
GPT-Engineer is BLOWING UP on GitHub right now 🔥
Prompt an AI agent to write an entire codebase 🤯
+2k stars today alone. I'll be experimenting with it this weekend to improve my workflow.
CJ Carr retweeted
Absolutely fantastic work. Congrats! \m/
Darius's "Hyperbolic Audio Source Separation" paper has been recognized as one of the top 3% of all papers accepted at #ICASSP2023 @ieeeICASSP 👏🎉🍾
Paper: arxiv.org/abs/2212.05008
Code: github.com/merlresearch/hype…
Demo video: youtube.com/watch?v=RKsAMb9z…
CJ Carr retweeted
Interesting alternative derivation of diffusion models without differential equations or variational inference. Reminiscent of flow matching / rectified flow.
Which of these perspectives is the simplest is subjective IMO, but more is better: new perspectives inspire new ideas!
When @_Laurent and I started learning about diffusion models, we were puzzled by the amount of jargon and concepts.
So, we derived a model from scratch with our own graphics-people intuitions. Simple derivation, simple implementation, SOTA quality.
ggx-research.github.io/publi…
CJ Carr retweeted
the year is 2030. after ed sheeran lost his case, musicians exhausted every 4-bar chord progression. pop music now features 23-bar cycles. tutorial videos explain how to escape copyright infringement using arabic 17 tone equal temperament. noise music experiences a renaissance.
CJ Carr retweeted
embodied experience as to what it is like to be a big bird
VRヘッドトラッキングを使ってドローンのパイロット気分を味わう
youtu.be/N5cBKmFZ6Lo
#FPV #POV #drone #OpenSource #HeadTracker #HMD #VR #VirtualReality #UX #experience
CJ Carr retweeted
Announcing StableLM❗
We’re releasing the first of our large language models, starting with 3B and 7B param models, with 15-65B to follow. Our LLMs are released under CC BY-SA license.
We’re also releasing RLHF-tuned models for research use. Read more→ stability.ai/blog/stability-…
CJ Carr retweeted
Rather than sign some letter to delay AI progress, I signed @LAION_AI petition to democratize AI research by establishing an international, publicly funded supercomputing facility equipped with 100K SOTA accelerators to train open source foundation models.
openpetition.eu/petition/onl…
CJ Carr retweeted
possibly one of the funniest model on the hub (the RLHF tutorial that was the reason of its creation is also awesome - check it out)
Excited to introduce: StackLlama🦙
An end-to-end tutorial for training Llama with RLHF on preference data such as the StackExchange questions!
Blog: hf.co/blog/stackllama
Demo: hf.co/spaces/trl-lib/stack-l…
Code: github.com/lvwerra/trl/tree/…
The resulting model is surprisingly fun!🧵
CJ Carr retweeted
Semi-realtime Stable Diffusion image-to-image (Van Gogh style),
Unity plugin based on Apple's Core ML port.
The latency is about 2 seconds on M1 Max MacBook Pro.
github.com/keijiro/StableDif…
CJ Carr retweeted
🎧🧠
🎵Introducing Diffusion Radio - A 24/7 Livestream of AI-Generated Music!🎵
📢The Harmonai team is excited to announce a New YouTube Live Stream which will run 24/7, showcasing amazing samples of AI-generated music from our latest models!📢
🎉Enjoy!🎉
youtube.com/watch?v=uGRLOMf2…
CJ Carr retweeted
New paper from our research team: First true multimodal image synthesis. You can prompt any combination of texts and images. Also shows the power of our multimodal embedding. Check out the Catronaut in the paper.
