abstract_datum's profile picture. Just data stuff all day

Abstract Data

@abstract_datum

Just data stuff all day

Abstract Data さんがリポスト

What an amazing week for OCR Models. Lord and Saviour @VikParuchuri has shipped an OCR model that outputs markdown, HTML, and JSON. This year was a blessing for Vision use cases, as we can be picky now. You know it's legit cause he is also named Vik huggingface.co/datalab-to/cha…


Abstract Data さんがリポスト

We launched a sandbox within Claude Code that allows you to define exactly which directories and network hosts your agent can access. Type /sandbox to enable it.

trq212's tweet image. We launched a sandbox within Claude Code that allows you to define exactly which directories and network hosts your agent can access.

Type /sandbox to enable it.

Abstract Data さんがリポスト

Man, this looks beautiful 😍 🧲 superseedr — A full-featured BitTorrent client for the terminal. 🌐 Supports multi-torrent downloads, DHT, PEX, magnet links & rarest-first piece selection. 🦀 Written in Rust & built with @ratatui_rs ⭐ GitHub: github.com/Jagalite/super…


Abstract Data さんがリポスト

DeepSeek OCR processes PDFs at 1/10th the cost of traditional OCR tools We're now hosting it on alphaXiv's API! Extract figures, complex diagrams, and text from any PDF 🚀


Abstract Data さんがリポスト

OminiControl Minimal and Universal Control for Diffusion Transformer

_akhaliq's tweet image. OminiControl

Minimal and Universal Control for Diffusion Transformer

Abstract Data さんがリポスト

New method for automatic prompt optimization! TL;DR: We do reinforcement learning to train an LM that can take in a lot of task examples and generate a prompt that describes the task. Trained/tested on 3000+ classification datasets on Hugging Face!

Can we train LLMs to be good prompt engineers? 🚀We propose Prompt-MII: Meta-Learning Instruction Induction for LLMs Our models out-perform strong baselines like ICL and GEPA with 13x fewer tokens. 🧵

XiaoEmily41333's tweet image. Can we train LLMs to be good prompt engineers?

🚀We propose Prompt-MII: Meta-Learning Instruction Induction for LLMs

Our models out-perform strong baselines like ICL and GEPA with 13x fewer tokens. 🧵


Abstract Data さんがリポスト

16000 drones over Liuyang, a new world record. We need to accelerate, quick. Especially europe.


Abstract Data さんがリポスト

🍁 Try this Black ink Sketch Art 🍁 Use Google Gemini ai 🍁 Send your edit result in comment section 🍁 Prompt 👇 Hyper-realistic monochrome ink 3D illustration — a 99.99% accurate portrait from the reference photo, drawn with black ink on textured white sketch paper.…

ShreyaYadav___'s tweet image. 🍁 Try this Black ink Sketch Art 🍁

Use Google Gemini ai 🍁

Send your edit result in comment section 🍁

Prompt 👇 

Hyper-realistic monochrome ink 3D illustration — a 99.99% accurate portrait from the reference photo, drawn with black ink on textured white sketch paper.…

Abstract Data さんがリポスト

I am incredibly excited to introduce rLLM v0.2. Zooming back to a year ago: @OpenAI's o1-preview just dropped, and RL + test-time scaling suddenly became the hype. But no one knew how they did it. @kylepmont and I had this idea - what if we built a solver-critique loop for…

🚀 Introducing rLLM v0.2 - train arbitrary agentic programs with RL, with minimal code changes. Most RL training systems adopt the agent-environment abstraction. But what about complex workflows? Think solver-critique pairs collaborating, or planner agents orchestrating multiple…

rllm_project's tweet image. 🚀 Introducing rLLM v0.2 - train arbitrary agentic programs with RL, with minimal code changes.

Most RL training systems adopt the agent-environment abstraction. But what about complex workflows? Think solver-critique pairs collaborating, or planner agents orchestrating multiple…


Abstract Data さんがリポスト

TIL you can just run `networkquality` in your terminal on MacOS

aidenybai's tweet image. TIL you can just run `networkquality` in your terminal on MacOS

Abstract Data さんがリポスト

shit is about to get so fucked up

justalexoki's tweet image. shit is about to get so fucked up

Abstract Data さんがリポスト

I'm a vibe coder

amikalash's tweet image. I'm a vibe coder

Abstract Data さんがリポスト

macOS Sandboxes now in Early Access on @daytonaio 💻

Launch Week 8, Day 5: Computer Use – macOS (Early Access)

daytonaio's tweet image. Launch Week 8, Day 5:  Computer Use – macOS (Early Access)


Abstract Data さんがリポスト

Smaller than your fingertip. Smarter than you’d expect. Yeah, it’s a full-blown RP2350 microcontroller with buttons, RGB, and HID power. HackBEE is coming. 🔥

hacktheboard's tweet image. Smaller than your fingertip. Smarter than you’d expect. Yeah, it’s a full-blown RP2350 microcontroller with buttons, RGB, and HID power.
 
HackBEE is coming. 🔥

Abstract Data さんがリポスト

AI bubble ended in the funniest way: 22: “AGI soon” 23: “AGI soon” 24: “AGI soon” 25: nvm we give up it’s just porn and ads


Abstract Data さんがリポスト

They're so back 👀 @mixedbreadai released two absolutely tiny late-interaction models for search: 17M and 32M parameters (that's 0.017B and 0.032B for those stuck thinking in B's 😆) Sequence length: 32k Integrated with PyLate Beats ColBERT-v2 on BEIR License: Apache 2.0 🧵


Abstract Data さんがリポスト

Our new Omni-Embed-Nemotron cross-modal retrieval model can search text, images, audio, or video using a single bi-encoder architecture that process all modalities in a shared embedding space for simpler, faster retrieval-- all in one open-source architecture 🛠️🌐 Try out the…

NVIDIAAIDev's tweet image. Our new Omni-Embed-Nemotron cross-modal retrieval model can search text, images, audio, or video using a single bi-encoder architecture that process all modalities in a shared embedding space for simpler, faster retrieval-- all in one open-source architecture 🛠️🌐

Try out the…

Abstract Data さんがリポスト

Isolates image subjects by removing backgrounds

tom_doerr's tweet image. Isolates image subjects by removing backgrounds

Abstract Data さんがリポスト

🚨PromptShare🚨 PENCIL DRAWING A pencil drawing of [character or object] [breaking through / emerging from / interacting with] [a paper surface or cracked wall], in the style of a tattoo sketch on white paper. Black pen and pencil only, with [one specific element] in [a vivid…

CharaspowerAI's tweet image. 🚨PromptShare🚨
PENCIL DRAWING

A pencil drawing of [character or object] [breaking through / emerging from / interacting with] [a paper surface or cracked wall], in the style of a tattoo sketch on white paper. Black pen and pencil only, with [one specific element] in [a vivid…
CharaspowerAI's tweet image. 🚨PromptShare🚨
PENCIL DRAWING

A pencil drawing of [character or object] [breaking through / emerging from / interacting with] [a paper surface or cracked wall], in the style of a tattoo sketch on white paper. Black pen and pencil only, with [one specific element] in [a vivid…
CharaspowerAI's tweet image. 🚨PromptShare🚨
PENCIL DRAWING

A pencil drawing of [character or object] [breaking through / emerging from / interacting with] [a paper surface or cracked wall], in the style of a tattoo sketch on white paper. Black pen and pencil only, with [one specific element] in [a vivid…
CharaspowerAI's tweet image. 🚨PromptShare🚨
PENCIL DRAWING

A pencil drawing of [character or object] [breaking through / emerging from / interacting with] [a paper surface or cracked wall], in the style of a tattoo sketch on white paper. Black pen and pencil only, with [one specific element] in [a vivid…

Abstract Data さんがリポスト

SQLite has a page where they explain why they use C. They specifically elaborate on why not Rust

iavins's tweet image. SQLite has a page where they explain why they use C. They specifically elaborate on why not Rust

Loading...

Something went wrong.


Something went wrong.