cs_peter_lorenz's profile picture. Post-Doc @NTUsg
Past: @UniHeidelberg @OptML_MSU @MITIBMLab @TUGraz

https://www.admonymous.co/peter-lorenz

Peter Lorenz

@cs_peter_lorenz

Post-Doc @NTUsg Past: @UniHeidelberg @OptML_MSU @MITIBMLab @TUGraz https://www.admonymous.co/peter-lorenz

Negotiating a job offer, and you don't know what the salary was in the past? Then, look up older job posts with web.archive.org


Exploring agents and LLMs? 🧠 Just came across a GitHub repo that looks like a solid foundation for anyone building in this space: 🔗 lnkd.in/gnYpFTHS Definitely one to bookmark and revisit. Credit to Hesam Sheikh Hassani for surfacing this on X! #LLM #Agents


🚀 Linux + Microsoft Edge = Cloud Powerhouse! 🐧💻 Just made a game-changing discovery on my Ubuntu machine: Microsoft Edge is the ultimate gateway to all my Microsoft cloud tools—CoPilot,Teams, Outlook, OneDrive, and more! 🌐✨


Repost di Peter Lorenz

🚀🚀Excited to share our paper R-Zero: Self-Evolving Reasoning LLM from Zero Data ! How to train LLM without data? R-Zero teaches Large Language Models to reason starting with nothing but a base model. No data required!!! Paper: arxiv.org/abs/2508.05004 Code:…

ChengsongH31219's tweet image. 🚀🚀Excited to share our paper R-Zero: Self-Evolving Reasoning LLM from Zero Data !

How to train LLM without data?

R-Zero teaches Large Language Models to reason starting with nothing but a base model. 
No data required!!!
Paper: arxiv.org/abs/2508.05004
Code:…

After daily using deepseek. I asked deepseek: "Generate an image of what it’s like to chat with me. Be honest and brutal." It is the best self reflection tool!


How would you start if you want to learn LLM training from scratch? Are there any sources that give you a good overview and trade-offs in a short time?


Repost di Peter Lorenz

I tried TRL again... I'm going back to OAT. Every time I try to use TRL, its always a nightmare. OAT is plug and play. github.com/sail-sg/oat


AI Agents = LLMs + memory + tools + autonomy Next-gen AI won’t just chat-it will do. Key ingredients: • Reasoning (LLMs) • Memory (RAG/vectors) • Action (APIs/tools) • Learning (RL) The future is AI that thinks and acts. #AI #AGI


I am going to ICLR in Singapore in a few weeks. I am happy to connect with new people. Let's meet for a coffee ☕ #ICLR2025 #ICLR #multimodality #computervision #trustworthy #AI


Agentic design patterns (in tech) 😃 What else am I missing here? What about Tiktok?

cs_peter_lorenz's tweet image. Agentic design patterns (in tech) 😃

What else am I missing here? What about Tiktok?

Never thought I'd be saying this but switching over to uv from pip was worth it :) I never thought that installing packages is so fast! #uv #pip #rust


Just published an article on open-source AI library alternatives! 🚀 As ML gets more data-heavy, new libraries are stepping up to meet the demand. Check it out here: medium.com/p/standard-ai-… #AI #MachineLearning #OpenSource


No GPU? No problem! 🚀 LeetGPU is your online CUDA playground to code, learn, and compete – no powerful GPU needed. Perfect for mastering CUDA, testing ideas, or leveling up your GPU skills! 🖥️💡 Check it out: leetgpu.com #CUDA #Programming #GPU

leetgpu.com

LeetGPU - The GPU Programming Challenge Platform

Learn, compete, and master GPU programming.


Your life isn't going to magically get better. Nobody is coming to save you. Only you can do that. It might not happen overnight, or next week, or next month... But stay consistent, work hard, and you'll eventually transform your situation into your ideal one.


Hi, have a nice day and enjoy the sun 😁


Repost di Peter Lorenz

🚀Towards efficient Diffusion Transformers! 😆We are happy to introduce RAS, the first diffusion sampling strategy that allows for regional variability in sampling ratios, achieving up to 2x+ speedup! 🔌Training-free, plug and play! 💪Nice work with @MSFTResearch @YangYou1991


LLM controlled robots through adversarial reasoning arxiv.org/pdf/2502.01633


Repost di Peter Lorenz

The future is so multi-agentic. In this paper, multi-agent system tackles complex data science tasks through collaborative problem-solving. Phase-based workflow with debugging ensures robust data science automation 🎯 Original Problem: Data science tasks with tabular data…

rohanpaul_ai's tweet image. The future is so multi-agentic.

In this paper, multi-agent system tackles complex data science tasks through collaborative problem-solving.

Phase-based workflow with debugging ensures robust data science automation

🎯 Original Problem:

Data science tasks with tabular data…

Loading...

Something went wrong.


Something went wrong.