Might have been faster to write the tweet myself #huggingchat
AI
-
HuggingChat Open-Source Model Approaching ChatGPT Capabilities
By
–
HuggingChat is a chat interface to test open-source chat models, atm https://
huggingface.co/OpenAssistant/
oasst-sft-6-llama-30b-xor
… by OpenAssistant. Agree that it's still a bit behind current ChatGPT but wouldn't be surprised if it catches up soon with unique oss capabilities and advantages -
DataRobot 9.0: Value-Driven AI Platform Innovations
By
–
How does the DataRobot AI Platform 9.0 help organizations realize value? Our approach to #ValueDrivenAI helps teams clearly calculate and track #ML impact, makes deployment automation easier, and much more. Read more about these innovations in our latest technical blog:
-
Computer Vision Inflection Point: Large Pre-trained Vision Transformers
By
–
2/2 Also, here's a video of the livestream where I discuss video prompting in detail, including why I think computer vision is reaching an inflection point (because of large pre-trained vision transformers). Let me know what you think!
-
Visual Prompting Revolutionizes Computer Vision Model Development
By
–
1/2 Visual Prompting takes ideas from text prompting, which has transformed NLP, to computer vision. Via a simple "visual prompt", you can build a vision model in seconds. Try it out here: https://
app.landing.ai/public/visual_
prompting
… -
Berkeley AI Lecture: Generative AI Meets Copyright Law
By
–
Tomorrow / Wed at 12:10pm PST, 6th @UCBerkeley Lecture on AI: Law Professor and copyright expert @PamelaSamuelson
: Generative AI Meets Copyright Law. Berkeley AI Research Lab @berkeley_ai @citrisnews Livestreamed: http://
berkeley.edu/ai -
Hyena Architecture Challenges Attention Dominance in AI Research
By
–
Attention may not be all you need, at least not according to the authors of Hyena: @MichaelPoli6
, @Massastrello
, Eric Nguyen, @realDanFu
, @tri_dao
, Stephen Baccus, Yoshua Bengio, @StefanoErmon
, Christopher Ré @HazyResearch
, @Mila_Quebec
, @UMontreal
, @SnorkelAI . Hyena can handle -
LayerNAS Reframes Multi-Objective Neural Architecture Search Optimization
By
–
Introducing LayerNAS, an approach that reframes multi-objective neural architecture search within the framework of combinatorial optimization to greatly reduce search complexity, enabling more efficient discovery of high-performing models. Read more → https://t.co/BCn7MPLAsu pic.twitter.com/9J2xpPiHhU
— Google AI (@GoogleAI) 25 avril 2023Introducing LayerNAS, an approach that reframes multi-objective neural architecture search within the framework of combinatorial optimization to greatly reduce search complexity, enabling more efficient discovery of high-performing models. Read more → https://
goo.gle/3oJnGxP -
Open Assistant seeks feedback on chat interface model dataset
By
–
We'd love to get as much feedback, comments or questions as possible about: – The interface: https://
huggingface.co/spaces/hugging
chat/chat-ui/discussions
…
– The model developed by Open Assistant: https://
huggingface.co/OpenAssistant/
oasst-sft-6-llama-30b-xor
…
– The dataset developed by Open Assistant: https://
huggingface.co/datasets/OpenA
ssistant/oasst1
… A million thanks to all!