rkennezangue's profile picture. Build infra for training, finetuning and serving multimodal deep learning, VLM, LLM, IR, SEARCH at web scale.

Rostand

@rkennezangue

Build infra for training, finetuning and serving multimodal deep learning, VLM, LLM, IR, SEARCH at web scale.

Joined November 2023
Rostand Reposted

This is how dictators destroy free nations. They threaten those who speak against them with death. We cannot entrust our country and our freedom to a petty, vindictive, cruel, unstable man who wants to be a tyrant. #Womenwillnotbesilenced #VoteKamala

Trump on Liz Cheney: "Let's put her with a rifle standing there with 9 barrels shooting at her. Let's see how she feels about it. You know, when the guns are trained on her face."



Rostand Reposted

A century ago, Germany’s democracy was hijacked by a bombastic and charismatic tyrant — and in this five-minute video, you’ll learn how he did it. Why did I produce this? Because I believe that history is speaking to us, and all too often, those who refuse to learn from it are…


Rostand Reposted

this is the BEST vision language model I have ever tried! Aria is a new model by @rhymes_ai_: a 25.3B multimodal model that can take image/video inputs 🤩 They release the model with Apache-2.0 license and fine-tuning scripts as well 👏 I tested it extensively, keep reading to…


Rostand Reposted

Hello, Pittsburgh! Tune in as I take the stage to talk about the importance of electing @KamalaHarris and @Tim_Walz


Rostand Reposted

The new Realtime API with web crawling is mind-blowing! Talk in realtime with any website. Powered by the OpenAI Realtime API and @firecrawl_dev 🔥 Check it out:


Rostand Reposted

This is still one of my favourite things on the internet

bandrui0's tweet image. This is still one of my favourite things on the internet

Rostand Reposted

European Parliament member from Germany, Moritz Körner demolished Hungarian dictator Viktor Orban on the floor of the parliament. Orban the Toad is a Kremlin puppet wreaking havoc and making the EU parliament dysfunctional.


Rostand Reposted

"Your invasion of Ukraine was carried out in your own interests. Only yours. To expand your mafia state into a mafia empire. An empire built on corruption... Imperialism: I recognize it when I see it. And I will call it what it is." - UK Foreign Secretary David Lammy at the UN…


Rostand Reposted

Pete says it best. It's not the crimes. It's not the scandals. It's not paying off a porn star. These aren't the things that will ultimately make Trump lose. Nope. It's Project 2025. "It's the simple fact that they wrote down their own policies." #DumpTrumpsProject2025


Rostand Reposted

Traitor


Rostand Reposted
's tweet image.

Rostand Reposted

I’m sorry but the richest man on the planet casually telling a woman he’s not dating that he’s going to put a baby in her is so skin crawlingly creepy I don’t know where to start

ArmandDoma's tweet image. I’m sorry but the richest man on the planet casually telling a woman he’s not dating that he’s going to put a baby in her is so skin crawlingly creepy I don’t know where to start

Rostand Reposted

This is some predator shit


Rostand Reposted

Bookmark it.

cooltechtipz's tweet image. Bookmark it.

Rostand Reposted

I asked Chatgpt if they would hire someone who gave this answer. 😂😂 “Based on this answer, I would have concerns about hiring this person. The response is notably disorganized and lacks clarity, making it difficult to understand their proposed approach to lowering child care…


Rostand Reposted

F# example to serve ONNX vision models (e.g. phi-3 vision) for online and batch modes. Supports multiple parallel request if GPU hardware allows. github.com/fwaris/OnnxVis…


Rostand Reposted

Fuck yeah! Moshi by @kyutai_labs 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.…

reach_vb's tweet image. Fuck yeah! Moshi by @kyutai_labs 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.…

Rostand Reposted

Local Gemma: a package for running LARGE language models with SMALL amounts of memory 🤏 The Gemma-2 27b model typically requires 70GB GPU memory. With Local Gemma, this is just 5GB 🤯 Try it for yourself in 2-steps: pip install local-gemma local-gemma --preset memory_extreme

sanchitgandhi99's tweet image. Local Gemma: a package for running LARGE language models with SMALL amounts of memory 🤏

The Gemma-2 27b model typically requires 70GB GPU memory. With Local Gemma, this is just 5GB 🤯

Try it for yourself in 2-steps:

pip install local-gemma
local-gemma --preset memory_extreme

Rostand Reposted

Meet our “Tiny Giant.” Our 1B parameter model xLAM-1B is now the best micro model for function calling, outperforming models 7x its size, including GPT-3.5 & Claude. On-device agentic AI is here. #AIResearch #SLM #TinyButMighty Paper: arxiv.org/pdf/2406.18518 Github:…

SFResearch's tweet image. Meet our “Tiny Giant.” Our 1B parameter model xLAM-1B is now the best micro model for function calling, outperforming models 7x its size, including GPT-3.5 & Claude. On-device agentic AI is here. #AIResearch #SLM #TinyButMighty

Paper: arxiv.org/pdf/2406.18518
Github:…

United States Trends
Loading...

Something went wrong.


Something went wrong.