[shardformer] support pipeline for deepseek v3 and optimize lora save (#6188)

* [shardformer] support pipeline for deepseek v3

* [checkpointio] fix lora save

* [devops] update ci env

* [booster] optimize lora

* fix test

* fix test
This commit is contained in:
Hongxin Liu
2025-02-14 14:48:54 +08:00
committed by GitHub
parent ec73f1b5e2
commit 014837e725
21 changed files with 478 additions and 91 deletions

View File

@@ -437,9 +437,6 @@ class TorchFSDPModel(ModelWrapper):
super().__init__(module)
self.module = FSDP(module, *args, **kwargs)
def unwrap(self):
return self.module
class FSDPOptimizerWrapper(OptimizerWrapper):
def __init__(self, optimizer: Optimizer, model: nn.Module):