From 3651cf68770e5cb58bc2edcfb2530aa58c3ae24d Mon Sep 17 00:00:00 2001 From: root Date: Thu, 31 Jan 2019 10:53:16 +0000 Subject: [PATCH] refine train --- fluid/PaddleCV/image_classification/train.py | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/fluid/PaddleCV/image_classification/train.py b/fluid/PaddleCV/image_classification/train.py index 27567558..7ac49bf2 100644 --- a/fluid/PaddleCV/image_classification/train.py +++ b/fluid/PaddleCV/image_classification/train.py @@ -18,7 +18,6 @@ import utils from utils.learning_rate import cosine_decay from utils.fp16_utils import create_master_params_grads, master_param_to_train_param from utility import add_arguments, print_arguments -import paddle.fluid.profiler as profiler IMAGENET1000 = 1281167 #IMAGENET100 = 128660 @@ -112,7 +111,6 @@ def optimizer_setting(params): momentum_rate = params["momentum_rate"] end_lr = 0 total_step = int((total_images / batch_size) * num_epochs) - print("linear decay total steps: ", total_step) lr = fluid.layers.polynomial_decay( start_lr, total_step, end_lr, power=1) optimizer = fluid.optimizer.Momentum( @@ -316,9 +314,14 @@ def train(args): batch_id = 0 try: while True: + if pass_id == 0 and batch_id == 5: + profiler.start_profiler("GPU") + elif pass_id == 0 and batch_id == 10: + profiler.stop_profiler("total", "profile") t1 = time.time() loss, acc1, acc5, lr = train_exe.run( fetch_list=train_fetch_list) + t2 = time.time() period = t2 - t1 loss = np.mean(np.array(loss)) @@ -329,6 +332,7 @@ def train(args): train_info[2].append(acc5) lr = np.mean(np.array(lr)) train_time.append(period) + if batch_id % 10 == 0: print("Pass {0}, trainbatch {1}, loss {2}, \ acc1 {3}, acc5 {4}, lr{5}, time {6}" @@ -336,6 +340,9 @@ def train(args): lr, "%2.2f sec" % period)) sys.stdout.flush() batch_id += 1 + if batch_id == 31: + exit(0) + except fluid.core.EOFException: train_py_reader.reset() -- GitLab