Benjamin Muller
@ben_mlr
Research in AI. Focusing on scaling language models multi-modally & multilingually. Llama pretraining team @AIatMeta
So many exciting releases from FAIR @AIatMeta Super happy to see Spirit LM now open-sourced. Spirit LM unlocks expressive speech generation through interleaving speech-text training and phonetic(hubert)+pitch+style-specific tokenization. Available here: Weights:…
Open science is how we continue to push technology forward and today at Meta FAIR we’re sharing eight new AI research artifacts including new models, datasets and code to inspire innovation in the community. More in the video from @jpineau1. This work is another important step…
Recent LLMs (e.g. LLama 3 🦙) are increasingly good at Math. However, this progress is reserved for languages with large amounts of task-specific instruct-tuning data. In this work @AIatMeta (led by @LucasBandarkar ), we introduce a new model merging technique called **Layer…
Cross-lingual transfer can be as easy as swapping model layers between LLMs! 🔀 Our model merging method can compose math and language skills by swapping top&bottom layers from a SFT’d target language expert into a math expert without retraining arxiv.org/pdf/2410.01335 🧵: [1/3]
We ran Llama 4 Maverick through some HELM benchmarks. It is 1st on HELM capabilities (MMLU-Pro, GPQA, IFEval, WildBench, Omni-MATH), but… crfm.stanford.edu/helm/capabilit…
Today is the start of a new era of natively multimodal AI innovation. Today, we’re introducing the first Llama 4 models: Llama 4 Scout and Llama 4 Maverick — our most advanced models yet and the best in their class for multimodality. Llama 4 Scout • 17B-active-parameter model…
🚨 Diverse Preference Optimization (DivPO) 🚨 SOTA LLMs have model collapse🫠: they can't generate diverse creative writing or synthetic data 🎨 DivPO trains for both high reward & diversity, vastly improving variety with similar quality. Paper 📝: arxiv.org/abs/2501.18101 🧵below
We released new research - Byte Latent Transformer(BLT) BLT encodes bytes into dynamic patches using light-weight local models and processes them with a large latent transformer. Think of it as a transformer sandwich!
New from Meta FAIR — Byte Latent Transformer: Patches Scale Better Than Tokens introduces BLT, which for the first time, matches tokenization-based LLM performance at scale with significant improvements in inference efficiency & robustness. Paper ➡️ go.fb.me/w23lmz
New from Meta FAIR — Byte Latent Transformer: Patches Scale Better Than Tokens introduces BLT, which for the first time, matches tokenization-based LLM performance at scale with significant improvements in inference efficiency & robustness. Paper ➡️ go.fb.me/w23lmz
Groundbreaking scaling trends for Byte-level Language Modeling with the new BLT architecture 🚀 More insights in the thread 🧵
🚀 Introducing the Byte Latent Transformer (BLT) – An LLM architecture that scales better than Llama 3 using byte-patches instead of tokens 🤯 Paper 📄 dl.fbaipublicfiles.com/blt/BLT__Patch… Code 🛠️ github.com/facebookresear…
Congrats @aymericzzz and team on being live! Very exciting vision to build entire softwares with just a prompt
Excited to share more about our background, vision and where we're headed at @agemoai with @r1ddhi at @BusinessInsider 𝗢𝘂𝗿 𝘃𝗶𝘀𝗶𝗼𝗻 𝗶𝘀 𝘁𝗼 𝗲𝗻𝗮𝗯𝗹𝗲 𝗮𝗻𝘆𝗼𝗻𝗲 𝘁𝗼 𝗰𝗿𝗲𝗮𝘁𝗲 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲 – from an idea to fully deployed software. The critical path to…
🌍 I’ve always had a dream of making AI accessible to everyone, regardless of location or language. However, current open MLLMs often respond in English, even to non-English queries! 🚀 Introducing Pangea: A Fully Open Multilingual Multimodal LLM supporting 39 languages! 🌐✨…
Meta Spirit LM: open source language model that mixes text and speech.
Today we released Meta Spirit LM — our first open source multimodal language model that freely mixes text and speech. Many existing AI voice experiences today use ASR to techniques to process speech before synthesizing with an LLM to generate text — but these approaches…
OK here goes the "excited to share ...." post Want to know how to train a T2V model (with other amazing capabilities) that beats ALL prior work ?? Well we released a 90 page tech report with every detail 😊 ai.meta.com/research/movie…… Thanks to the amazing team!
🎥 Today we’re premiering Meta Movie Gen: the most advanced media foundation models to-date. Developed by AI research teams at Meta, Movie Gen delivers state-of-the-art results across a range of capabilities. We’re excited for the potential of this line of research to usher in…
Introducing *Transfusion* - a unified approach for training models that can generate both text and images. arxiv.org/pdf/2408.11039 Transfusion combines language modeling (next token prediction) with diffusion to train a single transformer over mixed-modality sequences. This…
LLM Evaluations are an important area of work — today we're announcing a new LLM Evaluation Research Grant to foster further innovation in this area. Recipients will get $200K in funding to support this work. We're accepting proposals until September 6 ➡️ go.fb.me/eym3xq
Starting today, open source is leading the way. Introducing Llama 3.1: Our most capable models yet. Today we’re releasing a collection of new Llama 3.1 models including our long awaited 405B. These models deliver improved reasoning capabilities, a larger 128K token context…
So… we trained a model and we wrote a paper about it. Have fun y’all! llama.meta.com/llama-download… ai.meta.com/research/publi…
I'm giving the opening Keynote at ICML 2024 on Tuesday the 23rd @ 9:30am CEST. I'll try empower folks to get Open Science back on track -- the free discussion of ideas is such an important aspect of AI progress, and we've been losing track. This is a complex topic, and I wont…
It was great to present the Spirit-LM model with @tuanh208 Spirit-LM is a foundation model that jointly learns text and expressive speech based on Llama 2. Thanks @twelve_labs for organizing the webinar Arxiv available here for more details: arxiv.org/abs/2402.05755
The recording of this webinar with @ben_mlr and @tuanh208 of @metaai is up! Watch here: youtu.be/oL5YoLNfdJM 📺 They discussed: - Challenges of expressive speech generation - SpiRit-LM combines TextLM and SpeechLM - Training recipe and generation samples - Can we observe the…
A restricted, safety aligned (no-image-out) version of Chameleon (7B/34B) is now open-weight! github.com/facebookresear… The team strongly believes in open-source. We had to do a lot of work to get this out to the public safely. Congrats to the Chameleon team!
I’m excited to announce our latest paper, introducing a family of early-fusion token-in token-out (gpt4o….), models capable of interleaved text and image understanding and generation. arxiv.org/abs/2405.09818