Upload pointrend_rcnn_R_50_FPN_3x_coco.yaml

PointRend Config

Reference: https://github.com/facebookresearch/detectron2/tree/main/projects/PointRend

pointrend_rcnn_R_50_FPN_3x_coco.yaml

@@ -0,0 +1,12 @@
+_BASE_: Base-PointRend-RCNN-FPN.yaml
+MODEL:
+  WEIGHTS: detectron2://ImageNetPretrained/MSRA/R-50.pkl
+  RESNETS:
+    DEPTH: 50
+SOLVER:
+  STEPS: (210000, 250000)
+  MAX_ITER: 270000
+# To add COCO AP evaluation against the higher-quality LVIS annotations.
+# DATASETS:
+#   TEST: ("coco_2017_val", "lvis_v0.5_val_cocofied")
+