[Feature] Support loading weights from ckpt engine worker (#11755)

Signed-off-by: Yang Kaiyong <yangkaiyong.yky@antgroup.com>
Signed-off-by: Cruz Zhao <CruzZhao@linux.alibaba.com>
Signed-off-by: Xuchun Shang <xuchun.shang@gmail.com>
Co-authored-by: Yang Kaiyong <yangkaiyong.yky@antgroup.com>
Co-authored-by: Cruz Zhao <CruzZhao@linux.alibaba.com>
Co-authored-by: Xuchun Shang <xuchun.shang@gmail.com>
Co-authored-by: Shangming Cai <csmthu@gmail.com>
This commit is contained in:
Teng Ma
2025-10-24 00:23:30 +08:00
committed by GitHub
parent b0b4f71679
commit 96a5e4dd79
15 changed files with 552 additions and 0 deletions

View File

@@ -1080,6 +1080,24 @@ class InitWeightsSendGroupForRemoteInstanceReqInput(BaseReq):
backend: str = "nccl"
# Now UpdateWeightsFromIPCReqInput and UpdateWeightsFromIPCReqOutput
# are only used by Checkpoint Engine (https://github.com/MoonshotAI/checkpoint-engine)
@dataclass
class UpdateWeightsFromIPCReqInput(BaseReq):
# ZMQ socket paths for each device UUID
zmq_handles: Dict[str, str]
# Whether to flush cache after weight update
flush_cache: bool = True
# Optional: Update weight version along with weights
weight_version: Optional[str] = None
@dataclass
class UpdateWeightsFromIPCReqOutput(BaseReq):
success: bool
message: str
@dataclass
class InitWeightsSendGroupForRemoteInstanceReqOutput(BaseReq):
success: bool