Penguin-VL is a compact vision-language model family built to study how far multimodal efficiency can be pushed by redesigning the vision encoder, rather than only scaling data or model size.

Speculative decoding can help AI chatbots improve throughput and reduce hardware demand by using a smaller model to draft tokens that a larger model validates.

Open-source OCR from Baidu eliminates the GPU memory wall that limits long-document parsing. Unlimited OCR uses a constant KV ...

Abstract: This paper proposes a decoder-only Transformer model for analyzing and clustering vehicle-to-vehicle interactions in highway merge zones. Leveraging deep learning, the model captures complex ...

The company said on Friday night that the federal government had ordered limits on its Mythos and Fable 5 A.I. systems, citing national security concerns. By Cade Metz and Dustin Volz Cade Metz ...

Target-speaker voice activity detection (TS-VAD) is a promising approach to speaker diarization. However, a comprehensive evaluation across diverse real-world datasets remains absent. In this work, we ...

The Minneapolis cafe removed prices from its menu in February and says the unconventional model is attracting support from the community google maps A Minnesota restaurant that stopped charging ...

As AI models get more powerful, there’s also going to be a level of danger. Anthropic announced Mythos not long ago as a model that was too dangerous for public availability, but a new version of that ...

Anthropic is bringing its most powerful AI model to the general public for the first time, but it’s doing it with guardrails. On Tuesday, the AI firm launched Claude Fable 5, the first publicly ...