AI Beyond Transformers

Last Week in AI

Because, in some sense, hallucination is all LLMs do. They are dream machines.

Andrej Karpathy

Hello Little Coders!

Most of the current AI systems are built on top of Transformers, Thanks to the Deep Neural Network Architecture that was originally introduced by Google. While we have already seen a lot of great LLMs being built of this architecture, It’s got its own challenges and that’s anything beyond Transformers is kinda big deal.

That’s got more juice last week 👇🏾

AI Beyond Transformers

  1. Mamba: Linear-Time Sequence Modeling with Selective State Spaces Paper Code

    Try Mamba live here https://huggingface.co/spaces/reach-vb/mamba (Might not be too impressive)

  2. Paving the way to efficient architectures: StripedHyena-7B, open source models offering a glimpse into a world beyond Transformers Together Launch

New Open Models 

  1. Mistral 7B MoE Launch Tweet 

  2. Magicoder: Source Code Is All You Need https://arxiv.org/pdf/2312.02120.pdf https://github.com/ise-uiuc/magicoder

  3. MagicAnimate: Temporally Consistent Human Image Animation using Diffusion Model

     https://huggingface.co/spaces/zcxu-eric/magicanimate

  4. Video to Densepose https://github.com/Flode-Labs/vid2densepose

  5. Announcing Purple Llama: Towards open trust and safety in the new world of generative AI https://ai.meta.com/blog/purple-llama-open-trust-safety-generative-ai/

  6. Llama-Guard  https://huggingface.co/meta-llama/LlamaGuard-7b 

     

New Datasets

  1. Anthropic Dataset on Discriminative Prompts- https://huggingface.co/datasets/Anthropic/discrim-eval 

Great work from our subscriber!

  1. OpenML Guide (by our Sub) https://www.openmlguide.org/ 

AI Funding

  1. French AI start-up Mistral secures €2bn valuation https://www.ft.com/content/ea29ddf8-91cb-45e8-86a0-f501ab7ad9bb 

  2. Replicate raises $40 million Series B led by a16z https://replicate.com/blog/series-b 

  3. Announcing our $50M Series C to build superhuman Speech AI models https://www.assemblyai.com/blog/announcing-our-50m-series-c-to-build-superhuman-speech-ai-models/ 

General News

  1. The AI Alliance  https://thealliance.ai/ 

  2. Sam Altman is the Time’s CEO of the Year - https://time.com/6342827/ceo-of-the-year-2023-sam-altman/ 

  3. Liquid AI: A New Generation of AI Models from First Principles https://www.liquid.ai/blog/new-generation-of-ai-models-from-first-principles 

  4. Microsoft CoPilot - https://blogs.microsoft.com/blog/2023/12/05/celebrating-the-first-year-of-copilot-with-significant-new-innovations/ 

  5. MLX is an array framework for machine learning on Apple silicon, brought to you by Apple machine learning research. https://github.com/ml-explore/mlx 

  6. Welcome to the Gemini era https://deepmind.google/technologies/gemini/#introduction 

  7. Helen Toner talks about OpenAI Drama https://archive.ph/z1958#selection-4567.82-4567.101 

  8. Relightable Gaussian Codec Avatars https://shunsukesaito.github.io/rgca/ 

What are you reading this week? Let me know!

Have feedback or interesting project, Hit reply 🙂