new 9b text to video model on vbench?!
RESEARCH
-
LLaMA-Omni: Open-Source GPT-4o Alternative from China
By
–
Nice！Here is an opensource GPT-4o from China— LLaMA-Omni.https://t.co/sA1G0uCd0fhttps://t.co/L5jxJRZb5Jhttps://t.co/YiCyg9NiQo
— 机器之心 JIQIZHIXIN (@jiqizhixin) 25 septembre 2024
The work proposes LLaMA-Omni, a novel model architecture designed for low-latency and high-quality speech interaction with LLMs. https://t.co/MOtcT90yThNice！Here is an opensource GPT-4o from China— LLaMA-Omni. https://
arxiv.org/pdf/2409.06666 https://
github.com/ictnlp/LLaMA-O
mni
… https://
huggingface.co/ICTNLP/Llama-3
.1-8B-Omni
… The work proposes LLaMA-Omni, a novel model architecture designed for low-latency and high-quality speech interaction with LLMs. -
Making Text Embedders Few-Shot Learners with In-Context Learning
By
–
Making Text Embedders Few-Shot Learners discuss: https://
huggingface.co/papers/2409.15
700
… Large language models (LLMs) with decoder-only architectures demonstrate remarkable in-context learning (ICL) capabilities. This feature enables them to effectively handle both familiar and novel tasks by -
Gen2Act: Human Video Generation for Generalizable Robot Manipulation
By
–
Gen2Act
— AK (@_akhaliq) 25 septembre 2024
Human Video Generation in Novel Scenarios enables Generalizable Robot Manipulation
discuss: https://t.co/2lTOeBcQqz
How can robot manipulation policies generalize to novel tasks involving unseen object types and new motions? In this paper, we provide a solution in… pic.twitter.com/KttbmLmzO1Gen2Act Human Generation in Novel Scenarios enables Generalizable Robot Manipulation discuss: https://
huggingface.co/papers/2409.16
283
… How can robot manipulation policies generalize to novel tasks involving unseen object types and new motions? In this paper, we provide a solution in -
AI Researcher Urges Newsom to Veto California SB 1047
By
–
Thank you @GavinNewsom for supporting math education in California. I hope you do the same with AI and veto SB 1047. Please do not heed to the misinformation from celebrities and people with vested interests. As an AI researcher who has worked for two decades I strongly believe
-
OpenAI o1 models top Scale’s SEAL Leaderboards, early impressions
By
–
OpenAI’s o1-preview and o1-mini now top Scale’s SEAL Leaderboards! On the Scale blog, I recap benchmarks for the new o1 models and share some early impressions from my time using them. (Link in comment!)
-
LLM Development Enters Innovation Phase: Path to Proto-AGI
By
–
We are entering the 3rd phase of LLM Development. 1st phase was early tinkering, Transformer to GPT-3
2nd phase was scaling
3rd phase is an innovation phase: what breakthroughs beyond o1 get us to a new proto-AGI paradigm https://
x.com/a16z/status/18
38594324956864644/video/1
… -
Auto-generating ontologies for general purpose machine learning systems
By
–
Quick answer: yes. Long answer: for truly general purpose learning, it’s hard to define an ontology upfront so it would be fun to explore auto-generating ontologies (maybe with embedding clustering, etc.), but for specific use cases (which I’m working on now), having an ontology
-
Can LLM-as-Judge Replace Human Evaluators?
By
–
Human preference studies continue to be the gold standard when it comes to evaluating #LLMs But as models become more sophisticated, we asked ourselves: can LLM-as-a-Judge replace human evaluators? Read more in our blog https://
sambanova.ai/blog/can-llama
-405b-outperform-gpt4
… #GPT4 -
TMLR and ICLR announce research partnership collaboration
By
–
I'm afraid not. TMLR does have a somewhat similar partnership with ICLR however: https://
blog.iclr.cc/2024/08/22/icl
r2025-tmlr-partnership/
…