[pipeline] support fp32 for HybridPlugin/merge shardformer test and pipeline test into one file (#4354)

* add naive optimizer for 3DPlugin/refactor gpt2 shardformer test

* merge tests of PP/DP/TP combinations into one test file

* fix bug when sync grad for dp in HybridPlugin

* update supported precisions for 3DPlugin/fix bug when shifting tp_degree

* improve the passing of lazy_init

* modify lazy_init/use sync_shared_params
This commit is contained in:
Baizhou Zhang
2023-08-01 17:29:09 +08:00
committed by Hongxin Liu
parent f13954cd58
commit 0ceec8f9a9
8 changed files with 187 additions and 142 deletions

View File

@@ -134,7 +134,7 @@ class MixedPrecisionOptimizer(OptimizerWrapper):
working_param = self.master_to_working_map[p]
if p is working_param:
continue
if working_param.grad is None:
if working_param.grad is not None:
p.grad = working_param.grad.data.float()
working_param.grad = None
total_norm = self._compute_grad_norm()