okello

ML @DukeU

[AI at Meta @AIatMeta]

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 with 16 experts.
• Industry-leading context window of 10M tokens.
• Outperforms Gemma 3, Gemini 2.0 Flash-Lite and Mistral 3.1 across a broad range of widely accepted benchmarks.

Llama 4 Maverick
• 17B-active-parameter model with 128 experts.
• Best-in-class image grounding with the ability to align user prompts with relevant visual concepts and anchor model responses to regions in the image.
• Outperforms GPT-4o and Gemini 2.0 Flash across a broad range of widely accepted benchmarks.
• Achieves comparable results to DeepSeek v3 on reasoning and coding — at half the active parameters.
• Unparalleled performance-to-cost ratio with a chat version scoring ELO of 1417 on LMArena.

These models are our best yet thanks to distillation from Llama 4 Behemoth, our most powerful model yet. Llama 4 Behemoth is still in training and is currently seeing results that outperform GPT-4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro on STEM-focused benchmarks. We’re excited to share more details about it even while it’s still in flight.

Read more about the first Llama 4 models, including training and benchmarks ➡️ https://t.co/9G3QgVdCkB
Download Llama 4 ➡️ https://t.co/eVomRvEr0w

[okello @okello_mark]

@WesKambale @ngamita @UCC_ED @UCC_Official These are absolutely valid points and concerns. I skimmed through the comments, and discussing whether people understand LLMs, whether they are on Twitter, etc., made the important point you wanted to communicate vanish.

[kepano @kepano]

what an honor, and a wonderful reminder to write more love letters 🥹

[Jia-Bin Huang @jbhuang0604]

How to write the Introduction?

As a junior student, writing the introduction of a research paper is arguably the most daunting part of paper writing. 😱

Here is a simple template I find useful:

3 Figures 🖼️ + 5 Questions 🤔

[okello @okello_mark]

@goodfellow_ian All the best @goodfellow_ian and quick recovery!.

[okello @okello_mark]

@nowaraga Katesh was a commissioner customs not domestic taxes as suggested by the article.

[okello @okello_mark]

@ssmusoke @Airtel_Ug Nice to know you play Fortnite 😎. Next time we meet its Fortnite...

[okello @okello_mark]

@lornamariak 😎

[AI Pub @ai__pub]

Best of AI Twitter (Sept. 18-25):

- OpenAI releases Whisper (human-level speech-to-text)
- Adversarial, interactive deepfakes
- Twitter "gossip" on how OpenAI gathers GPT-4's trillions of training tokens
- The guts of Google's multibillion-parameter ad model

... and more:

1/15

[azul (she/her) @azulgarza_]

🎉 StatsForecast Exponential Smoothing (ETS) is 32% more accurate and over 100 times faster than NeuralProphet🔥

We hope this exercise by @nixtlainc helps the forecast community avoid adopting forecasting methods that are not thoroughly tested💡

#forecasting #python #timeseries

[Asianut @FidelOjulong]

Congratulations Sir
🎊 👏 💐

[Sayak Paul @RisingSayak]

.

@DeepMind released this GOLD a couple days back. If you ever wanted to study Transformers from scratch I think this would be that one resource you wouldn't want to miss:

https://t.co/ORtQAinlUj

[okello @okello_mark]

@UmemeLtd On a serious note, Pole replacement and conductor upgrade in this area if a serious contractor was awarded shouldn't take more than 2 days. Now its two months, Do you assess the work man hours wasted and its effects on the businesses. Sort your contracting!.

[benshi_ai @benshi_ai]

We are eager to participate in @phpulpocon'22 in September in #Vigo! Our contribution is twofold: @anafrio will deliver a talk, and we are supporting the conference through a gold sponsorship. Moitas ganas!
https://t.co/lSC0Y4WdU3

[okello @okello_mark]

@ngamita 😂

[okello @okello_mark]

@_jasonwei This is so brilliant. Thanks

[Jason Wei @_jasonwei]

In the spirit of the PhD admissions season ending, I'm making my state of purpose public.

I learned a lot from reading @nelsonfliu and @ssgrn's SoPs, and so I'd like to pay it forward.

https://t.co/awYteBgHow
https://t.co/MhJQc5nBnP

Prior SoPs below:

[Google AI @GoogleAI]

Introducing the Common Voice-based Speech-to-Speech translation corpus, CVSS, which includes 2657 hours of speech-to-speech translation sentence pairs from 21 languages into English. Read about its development and how we used it to train baseline models ↓ https://t.co/dxtif8v9bM

[Aidan Gomez @aidangomez]

Great article by @_RobToews on some of the major applications of NLP powering our interaction with technology.

https://t.co/G9VFiZ8VTR