Popular repositories Loading

1. FlashMLA FlashMLA Public

   Forked from deepseek-ai/FlashMLA

   C++

2. DeepSeek-V3 DeepSeek-V3 Public

   Forked from deepseek-ai/DeepSeek-V3

   Python

3. DeepEP DeepEP Public

   Forked from deepseek-ai/DeepEP

   DeepEP: an efficient expert-parallel communication library

   Cuda

4. DeepSeek-R1 DeepSeek-R1 Public

   Forked from deepseek-ai/DeepSeek-R1

5. DeepGEMM DeepGEMM Public

   Forked from deepseek-ai/DeepGEMM

   DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

   Cuda

6. DualPipe DualPipe Public

   Forked from deepseek-ai/DualPipe

   A bidirectional pipeline parallelism algorithm for computation-communication overlap in V3/R1 training.

   Python