NVIDIA GB300 NVL72 tokens per watt advantage over Hopper reaches 25x on DeepSeek V4 Pro per SemiAnalysis InferenceX data, ...

These local LLMs are changing the game in lots of fun ways.

For visual generation, discrete autoregressive models often struggle with poor tokenizer reconstruction, difficulties in sampling from large vocabularies, and slow token-by-token generation speeds. We ...

NVIDIA has released Nemotron-Labs-TwoTower, a diffusion language model built on a pretrained autoregressive backbone. It ships as open weights under the NVIDIA Nemotron Open Model License. The release ...

Autoregressive image generation aims to predict the next token based on previous ones. However, existing image tokenizers encode tokens with bidirectional dependencies during the compression process, ...

Abstract: The pre-training architectures of large language models encompass various types, including autoencoding models, autoregressive models, and encoder-decoder models. We posit that any modality ...

Abstract: Selective state space models (SSMs) represented by Mamba have demonstrated their computational efficiency and promising outcomes in various tasks, including automatic speech recognition (ASR ...

Large hybrid MoE models like Nemotron-3-Super are accurate but expensive to serve. Their active parameters, KV cache, and Mamba state cap how many users a node can hold at a given per-user token rate.