further clean

benchmark/fluid/args.py

@@ -136,10 +136,6 @@ def parse_args():
         '--no_random',
         action='store_true',
         help='If set, keep the random seed and do not shuffle the data.')
-    parser.add_argument(
-        '--use_lars',
-        action='store_true',
-        help='If set, use lars for optimizers, ONLY support resnet module.')
     parser.add_argument(
         '--reduce_strategy',
         type=str,

benchmark/fluid/models/resnet.py

@@ -200,11 +200,6 @@ def get_model(args, is_train, main_prog, startup_prog):
             # configure optimize
             optimizer = None
             if is_train:
-                if args.use_lars:
-                    lars_decay = 1.0
-                else:
-                    lars_decay = 0.0
-
                 total_images = 1281167 / trainer_count
 
                 step = int(total_images / (args.batch_size * args.gpus) + 1)

benchmark/fluid/models/resnet_with_preprocess.py

@@ -224,11 +224,6 @@ def get_model(args, is_train, main_prog, startup_prog):
             # configure optimize
             optimizer = None
             if is_train:
-                if args.use_lars:
-                    lars_decay = 1.0
-                else:
-                    lars_decay = 0.0
-
                 total_images = 1281167 / trainer_count
 
                 step = int(total_images / args.batch_size + 1)

benchmark/fluid/models/se_resnext.py

@@ -244,11 +244,6 @@ def get_model(args, is_train, main_prog, startup_prog):
 
             optimizer = None
             if is_train:
-                if args.use_lars:
-                    lars_decay = 1.0
-                else:
-                    lars_decay = 0.0
-
                 total_images = 1281167 / trainer_count
 
                 step = int(total_images / args.batch_size + 1)
@@ -262,8 +257,7 @@ def get_model(args, is_train, main_prog, startup_prog):
                     learning_rate=fluid.layers.piecewise_decay(
                         boundaries=bd, values=lr),
                     momentum=0.9,
-                    regularization=fluid.regularizer.L2Decay(1e-4),
-                    LARS_weight_decay=lars_decay)
+                    regularization=fluid.regularizer.L2Decay(1e-4))
                 optimizer.minimize(avg_cost)
 
                 if args.memory_optimize:

python/paddle/fluid/optimizer.py

@@ -104,7 +104,6 @@ class Optimizer(object):
         param = param_and_grad[0]
         param_lr = param.optimize_attr['learning_rate']
         if type(param_lr) == Variable:
-            # param learning rate has been updated (LARS)
             print("returns updated param lr ", param_lr)
             return param_lr
         else:

python/paddle/fluid/tests/book/test_recognize_digits.py

@@ -99,7 +99,7 @@ def train(nn_type,
 
     test_program = fluid.default_main_program().clone(for_test=True)
 
-    optimizer = fluid.optimizer.Adam(learning_rate=0.001, LARS_weight_decay=0.3)
+    optimizer = fluid.optimizer.Adam(learning_rate=0.001)
     optimizer.minimize(avg_loss)
 
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()