LLMS

→ View original post on X — @stabilityai

OpenFlamingo v2: Enhanced Multimodal AI Model Released

By

@stabilityai

–

29 June 2023 12h25

Our friends @anas_awadalla (U of Washington) and @irena_gao (Stanford) released this cool project with our compute support: OpenFlamingo v2: New Models and Enhanced Training Setup. Pls check it out on our AI blog and provide feedback!

Energy Efficiency in Open-Source Models at Scale

By

–

29 June 2023 0h28

How do you think about energy efficiency of training and inference at scale when open-source models are applied? Do you see low-energy models at the edge as entering the broader conversation about model-context match? @bmorphism

MosaicML’s LLM Development Philosophy: Aggressive vs Conservative Approach

By

–

29 June 2023 0h28

What is @MosaicML
's philosophical approach to developing LLMs? Is it like OpenAI: throw everything at the wall and see what sticks? More conservative, like Anthropic, Cohere, etc.? Or like FB, which open-sources everything and lets people figure out use cases? @PrajnaPrayas

Why Train Models on Multiple Languages Instead of Single Language

By

–

29 June 2023 0h28

Why do we need to train a model on 30 different languages? Why can't we train a model on just 1 language say python or JS ? Is it because there isn't enough training data ? – @lazyiitian

LLMs Deep Dive: Continuing Conversation Beyond Q&A

By

–

29 June 2023 0h28

We had such a great conversation with @jefrankle and couldn't get to the audience Q+A! Let's keep digging into LLMs by answering some of your questions!