4 Commits

Author	SHA1	Message	Date
lzy	d339df2e90	Supports DP+TP+EP hybrid parallel deployment strategy (#3489) ... * Support DP+TP+EP hybrid parallel deployment strategy * Support DP+TP+EP hybrid parallel deployment strategy * fix conflict * add moe_tp_ep function split_allgather_out * del tp_group in moe_cutlass_backend * for ci * fix parallel_config for ci * del log	2025-08-26 00:04:01 -07:00
Yuanle Liu	fac2f64837	delete parallel_state.py (#3250)	2025-08-08 11:03:29 +08:00
zhink	1272c7ce98	Fix performance degradation bug of custom_all_reduce (#2981)	2025-07-23 17:45:44 +08:00
zhink	0262ef7eb3	custom all reduce support cuda graph (#2938) ... * Support enabling cuda graph and custom all reduce at the same time, and fix the overwritten custom all reduce flag * rename communication_op to communication	2025-07-21 22:52:03 +08:00