add a field to support the evaluation interval (#849)

configs/mask_rcnn_r50_fpn_1x.py

@@ -171,6 +171,7 @@ log_config = dict(
         # dict(type='TensorboardLoggerHook')
     ])
 # yapf:enable
+evaluation = dict(interval=1)
 # runtime settings
 total_epochs = 12
 dist_params = dict(backend='nccl')

mmdet/apis/train.py

@@ -91,8 +91,8 @@ def build_optimizer(model, optimizer_cfg):
     paramwise_options = optimizer_cfg.pop('paramwise_options', None)
     # if no paramwise option is specified, just use the global setting
     if paramwise_options is None:
-        return obj_from_dict(
-            optimizer_cfg, torch.optim, dict(params=model.parameters()))
+        return obj_from_dict(optimizer_cfg, torch.optim,
+                             dict(params=model.parameters()))
     else:
         assert isinstance(paramwise_options, dict)
         # get base lr and weight decay
@@ -154,15 +154,19 @@ def _dist_train(model, dataset, cfg, validate=False):
     # register eval hooks
     if validate:
         val_dataset_cfg = cfg.data.val
+        eval_cfg = cfg.get('evaluation', {})
         if isinstance(model.module, RPN):
             # TODO: implement recall hooks for other datasets
-            runner.register_hook(CocoDistEvalRecallHook(val_dataset_cfg))
+            runner.register_hook(
+                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
         else:
             dataset_type = getattr(datasets, val_dataset_cfg.type)
             if issubclass(dataset_type, datasets.CocoDataset):
-                runner.register_hook(CocoDistEvalmAPHook(val_dataset_cfg))
+                runner.register_hook(
+                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
             else:
-                runner.register_hook(DistEvalmAPHook(val_dataset_cfg))
+                runner.register_hook(
+                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))
 
     if cfg.resume_from:
         runner.resume(cfg.resume_from)

mmdet/core/evaluation/eval_hooks.py

@@ -116,9 +116,11 @@ class CocoDistEvalRecallHook(DistEvalHook):
 
     def __init__(self,
                  dataset,
+                 interval=1,
                  proposal_nums=(100, 300, 1000),
                  iou_thrs=np.arange(0.5, 0.96, 0.05)):
-        super(CocoDistEvalRecallHook, self).__init__(dataset)
+        super(CocoDistEvalRecallHook, self).__init__(
+            dataset, interval=interval)
         self.proposal_nums = np.array(proposal_nums, dtype=np.int32)
         self.iou_thrs = np.array(iou_thrs, dtype=np.float32)