diff --git a/README.md b/README.md
index 8d34cf3..831d222 100644
--- a/README.md
+++ b/README.md
@@ -87,7 +87,7 @@ We followed STAC[1] to evaluate on 5 different data splits for each setting, and
 
 #### Notes
 - We use [wandb](https://wandb.ai/) for visualization, if you don't want to use it, just comment line `273-284` in `configs/soft_teacher/base.py`.
-
+- The project should be compatible to the latest version of `mmdetection`. If you want to switch to the same version `mmdetection` as ours, run `cd thirdparty/mmdetection && git checkout v2.16.0`
 ### Installation
 ```
 make install
diff --git a/configs/soft_teacher/base.py b/configs/soft_teacher/base.py
index 3b1978c..da907ac 100644
--- a/configs/soft_teacher/base.py
+++ b/configs/soft_teacher/base.py
@@ -183,7 +183,7 @@
     # generate fake labels for data format compatibility
     dict(type="PseudoSamples", with_bbox=True),
     dict(
-        type="MultiBranch", unsup_teacher=strong_pipeline, unsup_student=weak_pipeline
+        type="MultiBranch", unsup_student=strong_pipeline, unsup_teacher=weak_pipeline
     ),
 ]
 
@@ -244,7 +244,7 @@
         pseudo_label_initial_score_thr=0.5,
         rpn_pseudo_threshold=0.9,
         cls_pseudo_threshold=0.9,
-        reg_pseudo_threshold=0.01,
+        reg_pseudo_threshold=0.02,
         jitter_times=10,
         jitter_scale=0.06,
         min_pseduo_box_size=0,