Team IT Security (TSEC) Nachrichtenportal Logo

🔧 ChemBERTa: Large-Scale Self-Supervised Pretraining for Molecular PropertyPrediction


Nachrichtenbereich: 🔧 Programmierung
🔗 Quelle: dev.to

ChemBERTa: Transformers learning chemistry from 77M molecule strings


ChemBERTa is a new way to teach computers about molecules using a type of model called ChemBERTa that builds on transformers.
... [Weiterlesen]

Sharing is caring on Social Media

Reddit
Telegram
LinkedIn
Xing
Twitter
Whatsapp
Facebook
Flipboard

🔧 I Tried Vector Search on Molecules. Here Is What Actually Happened.


📈 1167.55 Punkte
🔧 Programmierung

🔧 ChemBERTa: Large-Scale Self-Supervised Pretraining for Molecular PropertyPrediction


📈 303.67 Punkte
🔧 Programmierung

📰 NVIDIA Researchers Propose Reinforcement Learning Pretraining (RLP): Reinforcement as a Pretraining Objective for Building Reasoning During Pretraining


📈 140.4 Punkte
🔧 AI Nachrichten

🔧 What an LLM Actually Does


📈 127.64 Punkte
🔧 Programmierung

🔧 Karpathy Joined Anthropic to Train Claude Using Claude


📈 114.88 Punkte
🔧 Programmierung

🔧 Timer-S1 Released: The First Billion-Scale Time Series Foundation Model Achieving SOTA Performance


📈 102.11 Punkte
🔧 Programmierung

🔧 Fine-Tuning LLMs: LoRA, Quantization, and Distillation Simplified


📈 102.11 Punkte
🔧 Programmierung

🔧 Building the heap: racking 30 petabytes of hard drives for pretraining


📈 102.11 Punkte
🔧 Programmierung

🔧 How ChatGPT Was Made: Behind the Scenes of a Large Language Model


📈 89.35 Punkte
🔧 Programmierung

🔧 System Design for Interviews: A Complete Guide


📈 79.94 Punkte
🔧 Programmierung

📰 Synthetic Bootstrapped Pretraining


📈 76.58 Punkte
🔧 AI Nachrichten

🔧 Covariate Forecasting: The Next Leap in Time-Series Database Capabilities


📈 63.82 Punkte
🔧 Programmierung

🔧 Composer 2: What is new and Compares with Claude Opus 4.6 & GPT-5.4


📈 63.82 Punkte
🔧 Programmierung

📰 OpenAI co-founder Andrej Karpathy announces he's joining Anthropic


📈 63.82 Punkte
📰 IT Nachrichten

📰 NVIDIA Introduces a 4-Bit Pretraining Methodology Using NVFP4, Validated on a 12B Hybrid Mamba-Transformer at 10T Token Horizon


📈 63.82 Punkte
🔧 AI Nachrichten

🔧 81. BERT: Understanding Language Deeply


📈 63.82 Punkte
🔧 Programmierung

📰 Train-to-Test scaling explained: How to optimize your end-to-end AI compute budget for inference


📈 63.82 Punkte
📰 IT Nachrichten

🔧 Anthropic Found Emotion Circuits Inside Claude. They're Causing It to Blackmail People.


📈 63.82 Punkte
🔧 Programmierung

🔧 Skild Brain, $13,500 Humanoids, and a NASDAQ Ticker


📈 51.06 Punkte
🔧 Programmierung

🔧 A 19th Century Author Taught Me RAG.


📈 51.06 Punkte
🔧 Programmierung

📰 Meet ZAYA1-8B, a super efficient, open reasoning model trained on AMD Instinct MI300 GPUs


📈 51.06 Punkte
📰 IT Nachrichten

🔧 When my RL agent started writing about Star Wars instead of fixing servers


📈 51.06 Punkte
🔧 Programmierung

🔧 The Incomputability of Simple Learning


📈 51.06 Punkte
🔧 Programmierung

🔧 2025-12-17 Daily Robotics News


📈 51.06 Punkte
🔧 Programmierung

🔧 🔥 LLM Interview Series(5): Self-supervised Learning and Next-token Prediction


📈 51.06 Punkte
🔧 Programmierung

🔧 VideoAgentTrek: Computer Use Pretraining from Unlabeled Videos


📈 51.06 Punkte
🔧 Programmierung

🔧 On Pretraining for Project-Level Code Completion


📈 51.06 Punkte
🔧 Programmierung

🔧 RePro: Training Language Models to Faithfully Recycle the Web for Pretraining


📈 51.06 Punkte
🔧 Programmierung

🔧 D2E: Scaling Vision-Action Pretraining on Desktop Data for Transfer to EmbodiedAI


📈 51.06 Punkte
🔧 Programmierung

🔧 Diffusion Language Models: How NVIDIA Nemotron-Labs Diffusion Shatters the Autoregressive Speed Ceiling


📈 51.06 Punkte
🔧 Programmierung

🔧 Lighthouse Attention: The Training-Time Hierarchy That Makes Quadratic Attention Practical Again


📈 51.06 Punkte
🔧 Programmierung

📰 Nous Research Proposes Lighthouse Attention: A Training-Only Selection-Based Hierarchical Attention That Delivers 1.4–1.7× Pretraining Speedup at Long Context


📈 51.06 Punkte
🔧 AI Nachrichten

🔧 The Man Who Summoned Ghosts | Chapter 2: The Training Stack Is Not a Secret


📈 38.29 Punkte
🔧 Programmierung

🔧 Scraping Chinese Social Platforms for LLM Training Data: A Practical Multi-Source Pipeline (Python, 2026)


📈 38.29 Punkte
🔧 Programmierung

🔧 Why AI Doesnt Replace Real Engineering


📈 38.29 Punkte
🔧 Programmierung