MiMo-V2.5
Xiaomi--moe-a2a-backend deepep
1 more from this model
--moe-a2a-backend deepep
implementation detail · filed under inference & serving
DeepEP is DeepSeek's library of all-to-all kernels that send each token to the GPUs holding its experts and gather the results back. Its first version had a high-throughput mode for training and prefill and a low-latency, CUDA-graph-capturable mode for decode; version 2 merged them into one interface, and version 2.5 removed the first.
Written for the explainer Kernels and graph compilation: where the cost model meets the GPU.
Also called --moe-a2a-backend deepep, DeepEP expert parallelism across nodes, DeepEP for expert parallelism.
4 quotes from 4 sources, one card per model, each opening with its strongest.
--moe-a2a-backend deepep
--moe-a2a-backend deepep
its SGLang example spans two nodes with --tp 16 --ep 16 and DeepEP for expert parallelism
--moe-a2a-backend deepep
Picked by hand, not extracted: where to read more, not evidence for anything on this page.