41. 🚀 DeepSeek-V3-0324 is out now! 🔹 Major boost in reasoning performance 🔹 Stronger front-end devel... 🚀 DeepSeek-V3-0324 is out now! 🔹 Major boost in reasoning performance 🔹 Stronger front-end development skills 🔹 Smarter tool-use capabilities ✅ For 03/25 21:32 ♡
42. & 🚀 Day 6 of #OpenSourceWeek: One More Thing – DeepSeek-V3/R1 Inference System Overview Optimized th... 🚀 Day 6 of #OpenSourceWeek k: One More Thing – DeepSeek-V3/R1 Inference System Overview Optimized throughput and latency via: 🔧 Cross-node EP-powered batch scaling 03/01 12:11 ♡
43. & 🚀 Day 5 of #OpenSourceWeek: 3FS, Thruster for All DeepSeek Data Access Fire-Flyer File System (3FS... 🚀 Day 5 of #OpenSourceWeek k: 3FS, Thruster for All DeepSeek Data Access Fire-Flyer File System (3FS) - a parallel file system that utilizes the full bandwidth of modern SS 02/28 09:06 ♡
44. & 🚀 Day 4 of #OpenSourceWeek: Optimized Parallelism Strategies ✅ DualPipe - a bidirectional pipeline... 🚀 Day 4 of #OpenSourceWeek k: Optimized Parallelism Strategies ✅ DualPipe - a bidirectional pipeline parallelism algorithm for computation-communication overlap in V3/R1 tr 02/27 10:05 ♡
45. & 🚀 Day 3 of #OpenSourceWeek: DeepGEMM Introducing DeepGEMM - an FP8 GEMM library that supports both... 🚀 Day 3 of #OpenSourceWeek k: DeepGEMM Introducing DeepGEMM - an FP8 GEMM library that supports both dense and MoE GEMMs, powering V3/R1 training and inference. ⚡ Up to 135 02/26 09:00 ♡
46. & 🚀 Day 2 of #OpenSourceWeek: DeepEP Excited to introduce DeepEP - the first open-source EP communic... 🚀 Day 2 of #OpenSourceWeek k: DeepEP Excited to introduce DeepEP - the first open-source EP communication library for MoE model training and inference. ✅ Efficient and opti 02/25 10:24 ♡
47. & 🚀 Day 1 of #OpenSourceWeek: FlashMLA Honored to share FlashMLA - our efficient MLA decoding kernel... 🚀 Day 1 of #OpenSourceWeek k: FlashMLA Honored to share FlashMLA - our efficient MLA decoding kernel for Hopper GPUs, optimized for variable-length sequences and now in pro 02/24 09:34 ♡
48. & 🚀 Day 0: Warming up for #OpenSourceWeek! We're a tiny team @deepseek_ai exploring AGI. Starting n... 🚀 Day 0: Warming up for #OpenSourceWeek k! We're a tiny team @deepseek_ai exploring AGI. Starting next week, we'll be open-sourcing 5 repos, sharing our small but s 02/21 12:00 ♡
49. 🚀 Introducing NSA: A Hardware-Aligned and Natively Trainable Sparse Attention mechanism for ultra-f... 🚀 Introducing NSA: A Hardware-Aligned and Natively Trainable Sparse Attention mechanism for ultra-fast long-context training & inference! Core components of NSA: • Dynamic 02/18 15:04 ♡
50. & 🎉 Excited to see everyone’s enthusiasm for deploying DeepSeek-R1! Here are our recommended settings... 🎉 Excited to see everyone’s enthusiasm for deploying DeepSeek-R1! Here are our recommended settings for the best experience: • No system prompt • Temperature: 0.6 • Officia 02/14 16:56 ♡
