mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2025-10-05 16:48:03 +08:00
[feat] support fa3 backend for pd disaggregated (#2695)
Some checks failed
Deploy GitHub Pages / deploy (push) Has been cancelled
Some checks failed
Deploy GitHub Pages / deploy (push) Has been cancelled
* support fa3 backend run in pd disaggregated * support fa3 backend run in pd disaggregated * support fa3 backend run in pd disaggregated * support fa3 backend run in pd disaggregated * delete use_fast_ffn
This commit is contained in:
@@ -86,7 +86,7 @@ class DefaultModelLoader(BaseModelLoader):
|
||||
if isinstance(v, paddle.Tensor):
|
||||
v.value().get_tensor()._clear()
|
||||
paddle.device.cuda.empty_cache()
|
||||
paddle.device.cuda.synchronize()
|
||||
paddle.device.synchronize()
|
||||
|
||||
def load_model(self, fd_config: FDConfig) -> nn.Layer:
|
||||
context = paddle.LazyGuard()
|
||||
|
Reference in New Issue
Block a user