Meta FAIR researchers have introduced JASCO, a generative text-to-music model. This new model differs from other text-to-music approaches through its ability to accept various conditioning inputs like chords or beat for improved controllability over the types of outputs it creates. You can find more details in the research paper available now, and we plan to release inference code and a pretrained model under a CC-BY-NC license soon. Details and samples ➡️ https://go.fb.me/tkp5fz
Looks great!
skip connexions and cross encoding during learning -> embed multimodal knowledge (ie. maps the knowledge with the information attached to its structured description or context -> bridges knowledge representation conversion and gives a map of language/low level logic statements) it is the concept of control nets (losely similar to), if I understand (?) , applied through learning, -> this set of encoder 'self training' to understand the very space of their own representation + continuous training -> possible logic escalation by meta cognition. (The multiple layers from the skip connexion providing the information needed to identify the depth/complexity of the specific feature's domain of representation ?) more prosaic, this goes well with discriminating between areas of knowledge representation (and somehow meets sparse encoding and model collapse's current works...provided I understood something ? )
Yann LeCun thats really cool. But Hey, I do disagree with you that we're not at artificial general intelligence, but I do agree that.It's not gonna come solely from language models.It's gonna require a multi model approach and a foundation model like gato or vima As well as the agent having a body system that uses a needs based statement machine so that it has motivations and a limbic system to guide its behavior
it is so much better than the slide show background pablum we have been subjected to online for wayyyy too long
For how long are we going to ignore the original audio/voices copyrights?
Meta FAIR's JASCO sounds fascinating! This generative text-to-music model stands out with its versatility!
This is so exciting!
Very cool! Love the choices! AI is not going to replace jobs, but the people that learn how to use AI, will take the jobs. “Hey Meta, what is the weather in San Jose right now?” Nice….
Love it! Seems like this might be the future of hyper-personalized remixes. #multimodal #genAI 🔊 🎶
AI Visionary and Aspiring Musician — Uncovering the Secrets of the Multiverse through AI Prompt Engineering, LLMs, and Music!👨🏼💻
2wInteresting 🤔 but the sound quality is still a far cry from being high quality and is a long way from other text-to-music AI tools like Suno… https://suno.com/song/4de0d22b-7919-4887-8602-659d888e8216