Add runner type

configs/_base_/schedules/schedule_160k.py

@@ -4,6 +4,6 @@
 # learning policy
 lr_config = dict(policy='poly', power=0.9, min_lr=1e-4, by_epoch=False)
 # runtime settings
-total_iters = 160000
+runner = dict(type='IterBasedRunner', max_iters=160000)
 checkpoint_config = dict(by_epoch=False, interval=16000)
 evaluation = dict(interval=16000, metric='mIoU')

configs/_base_/schedules/schedule_20k.py

@@ -4,6 +4,6 @@
 # learning policy
 lr_config = dict(policy='poly', power=0.9, min_lr=1e-4, by_epoch=False)
 # runtime settings
-total_iters = 20000
+runner = dict(type='IterBasedRunner', max_iters=20000)
 checkpoint_config = dict(by_epoch=False, interval=2000)
 evaluation = dict(interval=2000, metric='mIoU')

configs/_base_/schedules/schedule_40k.py

@@ -4,6 +4,6 @@
 # learning policy
 lr_config = dict(policy='poly', power=0.9, min_lr=1e-4, by_epoch=False)
 # runtime settings
-total_iters = 40000
+runner = dict(type='IterBasedRunner', max_iters=40000)
 checkpoint_config = dict(by_epoch=False, interval=4000)
 evaluation = dict(interval=4000, metric='mIoU')

configs/_base_/schedules/schedule_80k.py

@@ -4,6 +4,6 @@
 # learning policy
 lr_config = dict(policy='poly', power=0.9, min_lr=1e-4, by_epoch=False)
 # runtime settings
-total_iters = 80000
+runner = dict(type='IterBasedRunner', max_iters=80000)
 checkpoint_config = dict(by_epoch=False, interval=8000)
 evaluation = dict(interval=8000, metric='mIoU')

docs/config.md

@@ -225,7 +225,7 @@ dist_params = dict(backend='nccl')  # Parameters to setup distributed training,
 log_level = 'INFO'  # The level of logging.
 load_from = None  # load models as a pre-trained model from a given path. This will not resume training.
 resume_from = None  # Resume checkpoints from a given path, the training will be resumed from the iteration when the checkpoint's is saved.
-workflow = [('train', 1)]  # Workflow for runner. [('train', 1)] means there is only one workflow and the workflow named 'train' is executed once. The workflow trains the model by 40000 iterations according to the total_iters.
+workflow = [('train', 1)]  # Workflow for runner. [('train', 1)] means there is only one workflow and the workflow named 'train' is executed once. The workflow trains the model by 40000 iterations according to the `runner.max_iters`.
 cudnn_benchmark = True  # Whether use cudnn_benchmark to speed up, which is fast for fixed input size.
 optimizer = dict(  # Config used to build optimizer, support all the optimizers in PyTorch whose arguments are also the same as those in PyTorch
     type='SGD',  # Type of optimizers, refer to https://github.com/open-mmlab/mmcv/blob/master/mmcv/runner/optimizer/default_constructor.py#L13 for more details
@@ -238,7 +238,9 @@ lr_config = dict(
     power=0.9,  # The power of polynomial decay.
     min_lr=0.0001,  # The minimum learning rate to stable the training.
     by_epoch=False)  # Whethe count by epoch or not.
-total_iters = 40000  # Total number of iterations.
+runner = dict(
+    type='IterBasedRunner', # Type of runner to use (i.e. IterBasedRunner or EpochBasedRunner)
+    max_iters=40000) # Total number of iterations. For EpochBasedRunner use `max_epochs`
 checkpoint_config = dict(  # Config to set the checkpoint hook, Refer to https://github.com/open-mmlab/mmcv/blob/master/mmcv/runner/hooks/checkpoint.py for implementation.
     by_epoch=False,  # Whethe count by epoch or not.
     interval=4000)  # The save interval.

mmseg/__init__.py

@@ -2,7 +2,7 @@
 
 from .version import __version__, version_info
 
-MMCV_MIN = '1.1.2'
+MMCV_MIN = '1.1.4'
 MMCV_MAX = '1.2.0'
 
 

mmseg/apis/train.py

@@ -1,9 +1,10 @@
 import random
+import warnings
 
 import numpy as np
 import torch
 from mmcv.parallel import MMDataParallel, MMDistributedDataParallel
-from mmcv.runner import IterBasedRunner, build_optimizer
+from mmcv.runner import build_optimizer, build_runner
 
 from mmseg.core import DistEvalHook, EvalHook
 from mmseg.datasets import build_dataloader, build_dataset
@@ -70,13 +71,21 @@ def train_segmentor(model,
     # build runner
     optimizer = build_optimizer(model, cfg.optimizer)
 
-    runner = IterBasedRunner(
-        model=model,
-        batch_processor=None,
-        optimizer=optimizer,
-        work_dir=cfg.work_dir,
-        logger=logger,
-        meta=meta)
+    if cfg.get('runner') is None:
+        cfg.runner = {'type': 'IterBasedRunner', 'max_iters': cfg.total_iters}
+        warnings.warn(
+            'config is now expected to have a `runner` section, '
+            'please set `runner` in your config.', UserWarning)
+
+    runner = build_runner(
+        cfg.runner,
+        default_args=dict(
+            model=model,
+            batch_processor=None,
+            optimizer=optimizer,
+            work_dir=cfg.work_dir,
+            logger=logger,
+            meta=meta))
 
     # register hooks
     runner.register_training_hooks(cfg.lr_config, cfg.optimizer_config,
@@ -96,11 +105,12 @@ def train_segmentor(model,
             dist=distributed,
             shuffle=False)
         eval_cfg = cfg.get('evaluation', {})
+        eval_cfg['by_epoch'] = cfg.runner['type'] != 'IterBasedRunner'
         eval_hook = DistEvalHook if distributed else EvalHook
         runner.register_hook(eval_hook(val_dataloader, **eval_cfg))
 
     if cfg.resume_from:
         runner.resume(cfg.resume_from)
     elif cfg.load_from:
         runner.load_checkpoint(cfg.load_from)
-    runner.run(data_loaders, cfg.workflow, cfg.total_iters)
+    runner.run(data_loaders, cfg.workflow)

mmseg/core/evaluation/eval_hooks.py

@@ -12,17 +12,27 @@ class EvalHook(Hook):
         interval (int): Evaluation interval (by epochs). Default: 1.
     """
 
-    def __init__(self, dataloader, interval=1, **eval_kwargs):
+    def __init__(self, dataloader, interval=1, by_epoch=False, **eval_kwargs):
         if not isinstance(dataloader, DataLoader):
             raise TypeError('dataloader must be a pytorch DataLoader, but got '
                             f'{type(dataloader)}')
         self.dataloader = dataloader
         self.interval = interval
+        self.by_epoch = by_epoch
         self.eval_kwargs = eval_kwargs
 
     def after_train_iter(self, runner):
         """After train epoch hook."""
-        if not self.every_n_iters(runner, self.interval):
+        if self.by_epoch or not self.every_n_iters(runner, self.interval):
+            return
+        from mmseg.apis import single_gpu_test
+        runner.log_buffer.clear()
+        results = single_gpu_test(runner.model, self.dataloader, show=False)
+        self.evaluate(runner, results)
+
+    def after_train_epoch(self, runner):
+        """After train epoch hook."""
+        if not self.by_epoch or not self.every_n_epochs(runner, self.interval):
             return
         from mmseg.apis import single_gpu_test
         runner.log_buffer.clear()
@@ -54,6 +64,7 @@ def __init__(self,
                  dataloader,
                  interval=1,
                  gpu_collect=False,
+                 by_epoch=False,
                  **eval_kwargs):
         if not isinstance(dataloader, DataLoader):
             raise TypeError(
@@ -62,11 +73,27 @@ def __init__(self,
         self.dataloader = dataloader
         self.interval = interval
         self.gpu_collect = gpu_collect
+        self.by_epoch = by_epoch
         self.eval_kwargs = eval_kwargs
 
     def after_train_iter(self, runner):
         """After train epoch hook."""
-        if not self.every_n_iters(runner, self.interval):
+        if self.by_epoch or not self.every_n_iters(runner, self.interval):
+            return
+        from mmseg.apis import multi_gpu_test
+        runner.log_buffer.clear()
+        results = multi_gpu_test(
+            runner.model,
+            self.dataloader,
+            tmpdir=osp.join(runner.work_dir, '.eval_hook'),
+            gpu_collect=self.gpu_collect)
+        if runner.rank == 0:
+            print('\n')
+            self.evaluate(runner, results)
+
+    def after_train_epoch(self, runner):
+        """After train epoch hook."""
+        if not self.by_epoch or not self.every_n_epochs(runner, self.interval):
             return
         from mmseg.apis import multi_gpu_test
         runner.log_buffer.clear()

tests/test_eval_hook.py

@@ -38,7 +38,7 @@ def train_step(self, data_batch, optimizer):
         return dict(loss=loss)
 
 
-def test_eval_hook():
+def test_iter_eval_hook():
     with pytest.raises(TypeError):
         test_dataset = ExampleModel()
         data_loader = [
@@ -75,6 +75,43 @@ def test_eval_hook():
                                                  logger=runner.logger)
 
 
+def test_epoch_eval_hook():
+    with pytest.raises(TypeError):
+        test_dataset = ExampleModel()
+        data_loader = [
+            DataLoader(
+                test_dataset,
+                batch_size=1,
+                sampler=None,
+                num_worker=0,
+                shuffle=False)
+        ]
+        EvalHook(data_loader, by_epoch=True)
+
+    test_dataset = ExampleDataset()
+    test_dataset.evaluate = MagicMock(return_value=dict(test='success'))
+    loader = DataLoader(test_dataset, batch_size=1)
+    model = ExampleModel()
+    data_loader = DataLoader(
+        test_dataset, batch_size=1, sampler=None, num_workers=0, shuffle=False)
+    optim_cfg = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0005)
+    optimizer = obj_from_dict(optim_cfg, torch.optim,
+                              dict(params=model.parameters()))
+
+    # test EvalHook with interval
+    with tempfile.TemporaryDirectory() as tmpdir:
+        eval_hook = EvalHook(data_loader, by_epoch=True, interval=2)
+        runner = mmcv.runner.EpochBasedRunner(
+            model=model,
+            optimizer=optimizer,
+            work_dir=tmpdir,
+            logger=logging.getLogger())
+        runner.register_hook(eval_hook)
+        runner.run([loader], [('train', 1)], 2)
+        test_dataset.evaluate.assert_called_once_with([torch.tensor([1])],
+                                                      logger=runner.logger)
+
+
 def multi_gpu_test(model, data_loader, tmpdir=None, gpu_collect=False):
     results = single_gpu_test(model, data_loader)
     return results
@@ -116,3 +153,41 @@ def test_dist_eval_hook():
         runner.run([loader], [('train', 1)], 1)
         test_dataset.evaluate.assert_called_with([torch.tensor([1])],
                                                  logger=runner.logger)
+
+
+@patch('mmseg.apis.multi_gpu_test', multi_gpu_test)
+def test_dist_eval_hook_epoch():
+    with pytest.raises(TypeError):
+        test_dataset = ExampleModel()
+        data_loader = [
+            DataLoader(
+                test_dataset,
+                batch_size=1,
+                sampler=None,
+                num_worker=0,
+                shuffle=False)
+        ]
+        DistEvalHook(data_loader)
+
+    test_dataset = ExampleDataset()
+    test_dataset.evaluate = MagicMock(return_value=dict(test='success'))
+    loader = DataLoader(test_dataset, batch_size=1)
+    model = ExampleModel()
+    data_loader = DataLoader(
+        test_dataset, batch_size=1, sampler=None, num_workers=0, shuffle=False)
+    optim_cfg = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0005)
+    optimizer = obj_from_dict(optim_cfg, torch.optim,
+                              dict(params=model.parameters()))
+
+    # test DistEvalHook
+    with tempfile.TemporaryDirectory() as tmpdir:
+        eval_hook = DistEvalHook(data_loader, by_epoch=True, interval=2)
+        runner = mmcv.runner.EpochBasedRunner(
+            model=model,
+            optimizer=optimizer,
+            work_dir=tmpdir,
+            logger=logging.getLogger())
+        runner.register_hook(eval_hook)
+        runner.run([loader], [('train', 1)], 2)
+        test_dataset.evaluate.assert_called_with([torch.tensor([1])],
+                                                 logger=runner.logger)