add eval curve to metrics.json

Reviewed By: rbgirshick

Differential Revision: D18033952

fbshipit-source-id: b835578d53372980ba9f1918b1229224774cebd2

MODEL_ZOO.md

@@ -38,7 +38,8 @@ Unless otherwise noted, the following settings are used for all runs:
   Although 1x models are heavily under-trained, we provide some ResNet-50 models with the 1x (~12 COCO epochs)
   training schedule for comparison when doing quick research iteration.
 * The *model id* column is provided for ease of reference.
-  To check downloaded file integrity: any model on this page contains its md5 prefix in its file name.
+  To check downloaded file integrity, any model on this page contains its md5 prefix in its file name.
+	Each model also comes with a metrics file with all the training statistics and evaluation curves.
 
 #### ImageNet Pretrained Models
 

detectron2/evaluation/panoptic_evaluation.py

@@ -42,7 +42,6 @@ def __init__(self, dataset_name, output_dir):
         }
 
         self._predictions_json = os.path.join(output_dir, "predictions.json")
-        self._predictions_dir = os.path.join(output_dir, "predictions")
 
     def reset(self):
         self._predictions = []
@@ -95,10 +94,6 @@ def evaluate(self):
         gt_folder = self._metadata.panoptic_root
 
         with tempfile.TemporaryDirectory(prefix="panoptic_eval") as pred_dir:
-            if "://" not in self._predictions_dir:
-                pred_dir = self._predictions_dir
-                os.makedirs(pred_dir, exist_ok=True)
-
             logger.info("Writing all panoptic predictions to {} ...".format(pred_dir))
             for p in self._predictions:
                 with open(os.path.join(pred_dir, p["file_name"]), "wb") as f: