update

PaddleCV/image_classification/dist_train/dist_train.py

@@ -66,7 +66,6 @@ def parse_args():
     add_arg('split_var',          bool, True,               "Split params on pserver.")
     add_arg('async_mode',         bool, False,              "Async distributed training, only for pserver mode.")
     add_arg('reduce_strategy',    str,  "allreduce",        "Choose from reduce or allreduce.")
-    add_arg('skip_unbalanced_data', bool, False,            "Skip data not if data not balanced on nodes.")
     add_arg('enable_sequential_execution', bool, False,            "Skip data not if data not balanced on nodes.")
     #for dgc
     add_arg('enable_dgc', bool, False,            "Skip data not if data not balanced on nodes.")
@@ -85,13 +84,11 @@ def get_device_num():
         device_num = subprocess.check_output(['nvidia-smi', '-L']).decode().count('\n')
     return device_num
 
-def prepare_reader(is_train, pyreader, args, pass_id=0):
-    # NOTE: allways set reader infinite when nccl2 mode to balance data
-    # between ranks
-    is_infinite = (args.update_method == "nccl2")
+def prepare_reader(is_train, pyreader, args, pass_id=1):
+    # NOTE: always use infinite reader for dist training
     if is_train:
         reader = train(data_dir=args.data_dir, pass_id_as_seed=pass_id,
-                       infinite=is_infinite)
+                       infinite=True)
     else:
         reader = val(data_dir=args.data_dir)
     if is_train:
@@ -335,9 +332,8 @@ def train_parallel(args):
         num_samples = 0
         start_time = time.time()
         batch_id = 1
-        # use pass_id+1 as per pass global shuffle for distributed training
-        prepare_reader(True, train_pyreader, args, pass_id + 1)
-        train_pyreader.start()
+        if pass_id == 0:
+            train_pyreader.start()
         while True:
             try:
                 if batch_id % 30 == 0:
@@ -355,11 +351,10 @@ def train_parallel(args):
                 break
             num_samples += args.batch_size
             batch_id += 1
-            if (args.skip_unbalanced_data or args.update_method == "nccl2") and batch_id >= steps_per_pass:
+            if batch_id >= steps_per_pass:
                 break
 
         print_train_time(start_time, time.time(), num_samples)
-        train_pyreader.reset()
         if pass_id >= args.start_test_pass:
             if args.multi_batch_repeat > 1:
                 copyback_repeat_bn_params(train_prog)
@@ -375,6 +370,7 @@ def train_parallel(args):
             if not os.path.isdir(model_path):
                 os.makedirs(model_path)
             fluid.io.save_persistables(startup_exe, model_path, main_program=train_prog)
+    train_pyreader.reset()
     startup_exe.close()
     print("total train time: ", time.time() - over_all_start)
 

PaddleCV/image_classification/dist_train/run_mp_mode.sh

@@ -15,5 +15,7 @@ PADDLE_TRAINING_ROLE="TRAINER" \
 PADDLE_CURRENT_ENDPOINT="127.0.0.1:716${i}" \
 PADDLE_TRAINER_ID="${i}" \
 FLAGS_selected_gpus="${i}" \
-python dist_train.py --model $MODEL --update_method nccl2 --batch_size 32 --fp16 1 --scale_loss 8 &> logs/tr$i.log &
+python -u dist_train.py --model $MODEL --update_method nccl2 \
+--batch_size 32 \
+--fp16 0 --scale_loss 1 &> logs/tr$i.log &
 done

PaddleCV/image_classification/reader.py

@@ -12,7 +12,7 @@ np.random.seed(0)
 DATA_DIM = 224
 
 THREAD = 8
-BUF_SIZE = 102400
+BUF_SIZE = 1024
 
 DATA_DIR = 'data/ILSVRC2012'
 
@@ -131,7 +131,7 @@ def _reader_creator(file_list,
                     color_jitter=False,
                     rotate=False,
                     data_dir=DATA_DIR,
-                    pass_id_as_seed=0,
+                    pass_id_as_seed=1,
                     infinite=False):
     def reader():
         with open(file_list) as flist:
@@ -176,7 +176,7 @@ def _reader_creator(file_list,
     return paddle.reader.xmap_readers(mapper, reader, THREAD, BUF_SIZE)
 
 
-def train(data_dir=DATA_DIR, pass_id_as_seed=0, infinite=False):
+def train(data_dir=DATA_DIR, pass_id_as_seed=1, infinite=False):
     file_list = os.path.join(data_dir, 'train_list.txt')
     return _reader_creator(
         file_list,
@@ -185,7 +185,8 @@ def train(data_dir=DATA_DIR, pass_id_as_seed=0, infinite=False):
         color_jitter=False,
         rotate=False,
         data_dir=data_dir,
-        pass_id_as_seed=pass_id_as_seed)
+        pass_id_as_seed=pass_id_as_seed,
+        infinite=infinite)
 
 
 def val(data_dir=DATA_DIR):