diff --git a/tutorial/autofinetune.md b/tutorial/autofinetune.md index d402d3d02e2b4f43a139a4bb3b23d4aa225b325b..f36e3fcbf4443bc9e46e8dc080f0cff4fe5b8750 100644 --- a/tutorial/autofinetune.md +++ b/tutorial/autofinetune.md @@ -108,7 +108,7 @@ $ hub autofinetune train.py --param_file=hparam.yaml --cuda=['1','2'] --popsize= > `--evaluate_choice`: 可选,设置自动优化超参的评价效果方式,可选fulltrail和populationbased, 默认为populationbased -> `--tuning_strategy`: 可选,设置自动优化超参策略,可选hazero和pshe2,默认为pshe2 +> `--tuning_strategy`: 可选,设置自动优化超参算法,可选hazero和pshe2,默认为pshe2 **NOTE**: