deepseek-ai/Engram

Augments Transformers with N-gram memory for knowledge lookup

View on index · View in 3D Map

Related repos

• DeepSeek-V3, a Mixture-of-Experts language model with 671B parameters and 37B active per token, focused on math and coding tasks
• Not new, but still interesting
• Data analysis platform using large language models
• DeepSeek-R1 GitHub repo with model code and info
• n8n node for DeepSeek AI automations
• unified multimodal model for both understanding and generating images and text
• Memory system for AI assistants
• DeepEP — communication library for Mixture-of-Experts and expert parallelism, with all-to-all GPU kernels and low-precision support
• Model with sparse attention for long-context scenarios
• DeepClaude: Combines DeepSeek R1 and Claude 3.5 for AI reasoning and coding