Clarifai’s new reasoning engine makes AI models faster and less expensive

The AI platform Clarifai has launched a new reasoning engine that it claims will make running AI models twice as fast and 40% ...
IEEE

Zonotopic Distributed Fusion Over Binary Sensor Networks With Bit Rate Allocation: A Coding–Decoding Approach

Abstract: In this article, the zonotopic distributed fusion estimation problem is investigated for a class of general nonlinear systems over binary sensor networks subject to unknown-but-bounded (UBB) ...
IEEE

An Efficient ADMM-Based Penalized Decoding Algorithm for Binary LDPC Code

Abstract: LDPC decoding algorithm based on the alternating direction method of multipliers (ADMM) technique has increased much attentions because of its outstanding decoding performance. To enhancing ...
GitHub

VinePPO: Unlocking RL Potential For LLM Reasoning Through Refined Credit Assignment

Code for reproducing the results in the VinePPO paper. This codebase also provides performant implementation (leveraging vLLM as inference engine*) of popular RL and RL-free baselines (such as PPO, ...
GitHub

CoDe: Collaborative Decoding Makes Visual Auto-Regressive Modeling Efficient

1.7x Speedup and 0.5x memory consumption on ImageNet-256 generation. Top: original VAR-d30; Bottom: CoDe N=8. Speed measurement does not include vae decoder We propose Collaborative Decoding (CoDe), a ...