Thank you @_akhaliq for showcasing our paper on efficient video diffusion. We propose content-motion latent diffusion model – an efficient video model that utilizes pretrained image diffusion models and a low-dimensional motion latent representation for video generation
@animaanandkumar
-
Galore: Democratizing AI with Less Computing Power
By
–
While #GTC24 may feature even more powerful hardware we help you do more with less. Try our Galore on @huggingface #opensource #gpupoor #democratization
-
Austrian President Honors Postdoc’s Outstanding PhD Research
By
–
So proud to have my postdoc @julberner receive the highest honor in Austria for his PhD by the President of Austria!
-
FourCastNet: First High-Resolution AI Weather Forecasting Model
By
–
Very proud to see the FourCastNet model that I helped build at @nvidia in collaboration with @BerkeleyLab and other universities be highlighted in Jensen's keynote at #GTC24 – FourCastNet was the very first high-resolution #AI model to show competitive performance for weather… pic.twitter.com/SwkFDJSRtB
— Prof. Anima Anandkumar (@AnimaAnandkumar) 19 mars 2024Very proud to see the FourCastNet model that I helped build at @nvidia in collaboration with @BerkeleyLab and other universities be highlighted in Jensen's keynote at #GTC24 – FourCastNet was the very first high-resolution #AI model to show competitive performance for weather
-
ProteinDT and MoleculeSTM Featured in AI Research
By
–
Thank you for featuring ProteinDT and MoleculeSTM @Shengchao_Liu
-
Preserving Open Source Community: A Call to Action
By
–
I love the open source community! We should fight to preserve this at all costs!
-
Llama 7B Training on Single RTX 4090 GPU Memory Efficient
By
–
For the first time, we show that the Llama 7B LLM can be trained on a single consumer-grade GPU (RTX 4090) with only 24GB memory. This represents more than 82.5% reduction in memory for storing optimizer states during training. Training LLMs from scratch currently requires huge
-
Local Neural Operators Learn Multi-Resolution Feature Extraction
By
–
How do we capture local features across multiple resolutions? While standard convolutional layers work only on a fixed input-resolution, we design local neural operators that learn integral and differential kernels, and are principled ways to extend standard convolutions to
-
Generative AI Advances Protein-Ligand Binding Prediction
By
–
Thrilled that our work on generative AI for dynamic protein-ligand binding is now on the cover of @Nature @ZhuoranQ @wn8_nie @tfmiller3 https://
nature.com/articles/s4225
6-024-00792-z
… -
AI and Science Panel at National AI Advisory Committee
By
–
Join me and @JeffDean for a panel on AI +Science at National Artificial Intelligence Advisory Committee (NAIAC) Public Meeting. Register: https://
events.nist.gov/profile/form/i
ndex.cfm?PKformID=0x27760abcd
…
Agenda: https://
nist.gov/system/files/d
ocuments/2024/02/13/2024-02-22NAIACMeetingAgenda.pdf
…