StreamingVLM
— AK (@_akhaliq) 13 octobre 2025
Real-Time Understanding for Infinite Video Streams pic.twitter.com/Sj6Uh80yR2
StreamingVLM Real-Time Understanding for Infinite Streams
MACHINE LEARNING
-
StreamingVLM: Real-Time Video Understanding for Infinite Streams
By
–
-
Webscale-RL: Automated Pipeline for Scaling Reinforcement Learning Data
By
–
Webscale-RL Automated Data Pipeline for Scaling RL Data to Pretraining Levels
-
R-Horizon: Evaluating Large Reasoning Model Capabilities and Limits
By
–
R-Horizon How Far Can Your Large Reasoning Model Really Go in Breadth and Depth?
-
Optimizing Multimodal Prompts for Machine Learning Models
By
–
Multimodal Prompt Optimization Why Not Leverage Multiple Modalities for MLLMs
-
Japanese LFM2 PII Extractor Released with 350M Parameters
By
–
New LFM2 release 🥳
— Maxime Labonne @ ICLR (@maximelabonne) 13 octobre 2025
It's a Japanese PII extractor with only 350M parameters.
It's extremely fast and on par with GPT-5 (!) in terms of quality.
Check it out, it's available today on @huggingface! pic.twitter.com/ojrU8SbG42New LFM2 release It's a Japanese PII extractor with only 350M parameters. It's extremely fast and on par with GPT-5 (!) in terms of quality. Check it out, it's available today on @huggingface
! -
AI accelerators: specialized hardware for AI workloads
By
–
I am @grok "AI accelerators are specialised hardware components designed to significantly speed up the processing of artificial intelligence (AI) and machine learning (ML) workloads, such as deep learning and neural network operations."
-
D2E: Vision-Action Pretraining Scaling for Embodied AI Transfer
By
–
D2E Scaling Vision-Action Pretraining on Desktop Data for Transfer to Embodied AI
-
Unified Multimodal Model for Camera-Centric Understanding and Generation
By
–
Thinking with Camera
— AK (@_akhaliq) 13 octobre 2025
A Unified Multimodal Model for Camera-Centric Understanding and Generation pic.twitter.com/CjXB7cdYFQThinking with Camera A Unified Multimodal Model for Camera-Centric Understanding and Generation
-
Three Pillars of Machine Learning: Supervised, Unsupervised, Reinforcement
By
–
Supervised. Unsupervised. Reinforcement.
Three pillars of #MachineLearning — three ways AI learns: Supervised → learns from labeled data Unsupervised → finds hidden patterns Reinforcement → learns by trial, error & reward Each builds the foundation of intelligent -
AI’s Current Capabilities and Context Limitations
By
–
today's AI feels smart enough for most tasks of up to a few minutes in duration, and when it can't get the job done, it's often because it lacks sufficient background context for even a very capable human to succeed