🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

High-throughput Generative Inference of Large Language Models with a Single GPU abs: arxiv.org/abs/2303.06865 github: github.com/FMInference/Fl…

Promises kept! We're thrilled to unveil the first gift of our compute portfolio - the NVIDIA GPU H100 PCIe. Check out the official PR ⬇️ and stay tuned for more information to come 👉 ow.ly/T9ui50Nh1js

NVIDIA CEO Jensen Huang's keynote is a must-see for anyone who wants to stay up-to-date on the latest discoveries in AI. Tune in to #GTC23 to discover the #AI breakthroughs that are shaping our future--online on March 21, 8 a.m. PDT. nvda.ws/3LlBNT3

I'm also curious to see this. GPT-3 scored ~0 on ARC. I'd expect GPT-4 to at least solve the tasks that are analogous to common IQ problems (i.e. the trivial subset of the training set). That said, doubt it could do anything with the (more novel) evaluation test.

So many things to get out of this! - Project hf.co/bigcode - Blog hf.co/blog/starcoder - Generate code hf.co/spaces/bigcode… -Play with it hf.co/spaces/bigcode… - Assess reasoning hf.co/spaces/bigcode… - Explore dataset hf.co/spaces/bigcode…

MPT-7B: A New Standard for Open-Source, Commercially Usable LLMs - MPT-7B-StoryWriter-65k+, uses a context length of 65k tokens! - Introducing MPT-7B, the latest entry in MosaicML Foundation Series - MPT-7B is a transformer trained from scratch on 1T tokens of text and code

We just released Transformers' boldest feature: Transformers Agents. This removes the barrier of entry to machine learning Control 100,000+ HF models by talking to Transformers and Diffusers Fully multimodal agent: text, images, video, audio, docs...🌎 huggingface.co/docs/transform…

Excited to announce our multi-modal agent API that can automatically chain 100,000+ of HF models (stable diffusion, whisper, OpenAssistant,...) for text, audio, image, video, time-series,… based on your commands. All open-source so it can work locally! huggingface.co/docs/transform…

What's new in TensorFlow and Keras? 🤯 a LOT! 👏 The TensorFlow ecosystem is better than ever, giving developers new tools to access the next generation of ML. Learn all about it ➡️ goo.gle/io23_TFandKeras

New tutorial on keras.io: training a LLM from scratch on TPU with 🤗 Transformers keras.io/examples/nlp/m…

QLoRA: 4-bit finetuning of LLMs is here! With it comes Guanaco, a chatbot on a single GPU, achieving 99% ChatGPT performance on the Vicuna benchmark: Paper: arxiv.org/abs/2305.14314 Code+Demo: github.com/artidoro/qlora Samples: colab.research.google.com/drive/1kK6xasH… Colab: colab.research.google.com/drive/17XEqL1J…

Open-source LLMs are behind commercial models when it comes to context length. 🔠 OpenAI GPT-3.5 now has 16k, GPT-4 of 32k and Anthropic Claude up 100k💪🏻  For example, Meta LLaMa or Falcon have only 2k😔 Here are two amazing blog posts I found in the last week .🚀😍 🧵 1/3

I've joined Scaleway for that kind of moment !  Exciting times ahead! If you're interested to use H100 #GPU or DGX H100, reach out 👉 ow.ly/uRGA50P5Ywg ! #Scaleway #AI #DeepLearning

Unleash the power to train your #LLMs and largest #AI models : Meet Scaleway 's NVIDIA DGX #H100 #SuperPOD 🤖🔥🚀 1016 H100 with 400Gb/s InfiniBand and high performance low latency DDN storage ! Hosted in Paris in DC5 datacenter with super efficient adiabatic cooling 🌱💧♻️

🚀 [BREAKING NEWS] Our subsidiary Scaleway is beginning to unveil the program for its European #conference on #AI. We’re honoured to announce that Jensen Huang, NVIDIA founder and CEO, will speak at the end of the morning! To know more 👉iliad.fr/en/actualites/…

Exciting news for attendees! Jensen Huang, NVIDIA founder and CEO is speaking at #ai-PULSE on Nov 17th. Don’t miss the valuable insights from this renowned industry leader at Europe’s must-attend #AI conference.

Amazing demo of Moshi 🤖, kyutai real time voice AI 👏 Trained on our "nabu" NVIDIA Europe #H100 #superpod, demo powered by our Scaleway L4 #GPUs instances 🔥🚀

Fuck yeah! Moshi by kyutai just owned the stage! 🇪🇺/acc. Architecture 1. 7B Multimodal LM (speech in, speech out) 2. 2 channel I/O - Streaming LM constantly generates text tokens as well as audio codecs (tunable) 3. Achieves 160ms latency (with a Real-Time Factor of 2) 4.

The kyutai fully end-to-end audio model demo of today is a huge deal that many people missed in the room Mostly irrelevant are the facts that: - they come a few week after OpenAI ChatGPT-4o - the demo was less polished than the 4o one (in terms of voice quality, voice

...we continue with Aude Durand, Deputy CEO at Groupe iliad , addressing a clear message: "We're here to rally the European AI ecosystem"! This year, #aiPULSE 2024 centers on 3 pillars: Big, Efficient, and Open. 1️⃣ BIG: "In 2023, we led the way with 1,000 GPUs. Today, we've