support merge_all_iters_to_one_epoch to avoid deadlock when switching…

… epochs, eval last 10 epochs after training

pcdet/datasets/dataset.py

@@ -34,6 +34,8 @@ def __init__(self, dataset_cfg=None, class_names=None, training=True, root_path=
 
         self.grid_size = self.data_processor.grid_size
         self.voxel_size = self.data_processor.voxel_size
+        self.total_epochs = 0
+        self._merge_all_iters_to_one_epoch = False
 
     @property
     def mode(self):
@@ -65,6 +67,13 @@ def generate_prediction_dicts(batch_dict, pred_dicts, class_names, output_path=N
 
         """
 
+    def merge_all_iters_to_one_epoch(self, merge=True, epochs=None):
+        if merge:
+            self._merge_all_iters_to_one_epoch = True
+            self.total_epochs = epochs
+        else:
+            self._merge_all_iters_to_one_epoch = False
+
     def __len__(self):
         raise NotImplementedError
 

pcdet/datasets/kitti/kitti_dataset.py

@@ -330,10 +330,16 @@ def evaluation(self, det_annos, class_names, **kwargs):
         return ap_result_str, ap_dict
 
     def __len__(self):
+        if self._merge_all_iters_to_one_epoch:
+            return len(self.kitti_infos) * self.total_epochs
+
         return len(self.kitti_infos)
 
     def __getitem__(self, index):
         # index = 4
+        if self._merge_all_iters_to_one_epoch:
+            index = index % len(self.kitti_infos)
+
         info = copy.deepcopy(self.kitti_infos[index])
 
         sample_idx = info['point_cloud']['lidar_idx']

tools/test.py

@@ -96,8 +96,9 @@ def repeat_eval_ckpt(model, test_loader, args, eval_output_dir, logger, ckpt_dir
         cur_epoch_id, cur_ckpt = get_no_evaluated_ckpt(ckpt_dir, ckpt_record_file, args)
         if cur_epoch_id == -1 or int(float(cur_epoch_id)) < args.start_epoch:
             wait_second = 30
-            print('Wait %s seconds for next check (progress: %.1f / %d minutes): %s \r'
-                  % (wait_second, total_time * 1.0 / 60, args.max_waiting_mins, ckpt_dir), end='', flush=True)
+            if cfg.LOCAL_RANK == 0:
+                print('Wait %s seconds for next check (progress: %.1f / %d minutes): %s \r'
+                      % (wait_second, total_time * 1.0 / 60, args.max_waiting_mins, ckpt_dir), end='', flush=True)
             time.sleep(wait_second)
             total_time += 30
             if total_time > args.max_waiting_mins * 60 and (first_eval is False):

tools/train.py

@@ -9,7 +9,7 @@
 from train_utils.optimization import build_optimizer, build_scheduler
 from train_utils.train_utils import train_model
 import torch.distributed as dist
-
+from test import repeat_eval_ckpt
 from pathlib import Path
 import argparse
 import datetime
@@ -129,9 +129,8 @@ def main():
         model = nn.parallel.DistributedDataParallel(model, device_ids=[cfg.LOCAL_RANK % torch.cuda.device_count()])
     logger.info(model)
 
-    total_iters_each_epoch = len(train_loader) if not args.merge_all_iters_to_one_epoch else len(train_loader) // args.epochs
     lr_scheduler, lr_warmup_scheduler = build_scheduler(
-        optimizer, total_iters_each_epoch=total_iters_each_epoch, total_epochs=args.epochs,
+        optimizer, total_iters_each_epoch=len(train_loader), total_epochs=args.epochs,
         last_epoch=last_epoch, optim_cfg=cfg.OPTIMIZATION
     )
 
@@ -161,6 +160,26 @@ def main():
     logger.info('**********************End training %s/%s(%s)**********************\n\n\n'
                 % (cfg.EXP_GROUP_PATH, cfg.TAG, args.extra_tag))
 
+    logger.info('**********************Start evaluation %s/%s(%s)**********************' %
+                (cfg.EXP_GROUP_PATH, cfg.TAG, args.extra_tag))
+    test_set, test_loader, sampler = build_dataloader(
+        dataset_cfg=cfg.DATA_CONFIG,
+        class_names=cfg.CLASS_NAMES,
+        batch_size=args.batch_size,
+        dist=dist_train, workers=args.workers, logger=logger, training=False
+    )
+    eval_output_dir = output_dir / 'eval' / 'eval_with_train'
+    eval_output_dir.mkdir(parents=True, exist_ok=True)
+    args.start_epoch = max(args.epochs - 10, 0)  # Only evaluate the last 10 epochs
+
+    repeat_eval_ckpt(
+        model.module if dist_train else model,
+        test_loader, args, eval_output_dir, logger, ckpt_dir,
+        dist_test=dist_train
+    )
+    logger.info('**********************End evaluation %s/%s(%s)**********************' %
+                (cfg.EXP_GROUP_PATH, cfg.TAG, args.extra_tag))
+
 
 if __name__ == '__main__':
     main()