ResearchLatent Space

[AINews] DeepSeek v4.1-Flash: 763B-P8B-D16B novel causal Encoder–Decoder architecture with vision marks the Return of the Whale

#deepseek#ai#architecture#encoder-decoder#innovation

English

DeepSeek has introduced its new architecture, DeepSeek v4.1-Flash, featuring a novel causal Encoder-Decoder design with visual capabilities. Despite being behind some benchmarks, the model emphasizes the innovative use of context and aims to enhance efficiency and scalability in AI applications.

中文

DeepSeek推出了其新架构DeepSeek v4.1-Flash,具有新颖的因果编码器-解码器设计和视觉能力。尽管在一些基准测试中落后,但该模型强调创新的上下文使用,旨在提高AI应用的效率和可扩展性。