From 40a8659c39c479373f1e81b2b5d56198674160e8 Mon Sep 17 00:00:00 2001 From: sneaxiy Date: Tue, 8 Oct 2019 20:04:57 +0800 Subject: [PATCH] remove gc flags, test=develop --- PaddleCV/PaddleDetection/tools/eval.py | 13 ------------- PaddleCV/PaddleDetection/tools/infer.py | 13 ------------- PaddleCV/PaddleDetection/tools/train.py | 13 ------------- PaddleCV/PaddleGAN/cycle_gan/train.py | 8 +------- PaddleCV/PaddleGAN/scripts/run_SPADE.sh | 2 -- PaddleCV/PaddleVideo/models/bmn/README.md | 2 -- PaddleCV/PaddleVideo/models/bsn/README.md | 4 ---- PaddleCV/PaddleVideo/models/ctcn/README.md | 2 -- PaddleCV/PaddleVideo/models/tsm/README.md | 2 -- PaddleCV/PaddleVideo/models/tsn/README.md | 2 -- PaddleCV/PaddleVideo/run.sh | 2 -- PaddleCV/PaddleVideo/run_ce.sh | 2 -- PaddleCV/deeplabv3+/train.py | 6 ------ PaddleCV/face_detection/train.py | 15 --------------- .../image_classification/scripts/train/AlexNet.sh | 2 -- .../scripts/train/EfficientNetB0.sh | 2 -- .../scripts/train/GoogLeNet.sh | 2 -- .../scripts/train/InceptionV4.sh | 2 -- .../scripts/train/MobileNetV1.sh | 2 -- .../scripts/train/MobileNetV1_x0_25.sh | 2 -- .../scripts/train/MobileNetV1_x0_5.sh | 2 -- .../scripts/train/MobileNetV1_x0_75.sh | 2 -- .../scripts/train/MobileNetV2.sh | 2 -- .../scripts/train/MobileNetV2_x0_75.sh | 2 -- .../scripts/train/ResNeXt101_32x4d.sh | 2 -- .../scripts/train/ResNeXt101_64x4d.sh | 2 -- .../scripts/train/ResNeXt101_vd_32x4d.sh | 2 -- .../scripts/train/ResNeXt101_vd_64x4d.sh | 2 -- .../scripts/train/ResNeXt152_64x4d.sh | 2 -- .../scripts/train/ResNet101.sh | 2 -- .../scripts/train/ResNet101_vd.sh | 2 -- .../scripts/train/ResNet152.sh | 2 -- .../scripts/train/ResNet152_vd.sh | 2 -- .../scripts/train/ResNet18.sh | 2 -- .../scripts/train/ResNet18_vd.sh | 2 -- .../scripts/train/ResNet200_vd.sh | 2 -- .../scripts/train/ResNet34.sh | 2 -- .../scripts/train/ResNet34_vd.sh | 2 -- .../scripts/train/ResNet50.sh | 2 -- .../scripts/train/ResNet50_vc.sh | 2 -- .../scripts/train/ResNet50_vd.sh | 2 -- .../scripts/train/SENet154_vd.sh | 2 -- .../scripts/train/SE_ResNeXt101_32x4d.sh | 2 -- .../scripts/train/SE_ResNeXt50_32x4d.sh | 2 -- .../scripts/train/ShuffleNetV2.sh | 2 -- .../scripts/train/ShuffleNetV2_swish.sh | 2 -- .../image_classification/scripts/train/VGG11.sh | 2 -- .../image_classification/scripts/train/VGG13.sh | 2 -- .../image_classification/scripts/train/VGG16.sh | 2 -- .../image_classification/scripts/train/VGG19.sh | 2 -- PaddleCV/image_classification/train.py | 1 - PaddleCV/rcnn/train.py | 2 -- PaddleCV/ssd/train.py | 14 -------------- PaddleCV/yolov3/train.py | 2 -- PaddleNLP/Research/ACL2018-DAM/run.sh | 1 - PaddleNLP/Research/ACL2018-DAM/run_CPU.sh | 1 - .../Research/MRQA2019-BASELINE/run_finetuning.sh | 1 - .../Research/MRQA2019-BASELINE/run_predict.sh | 1 - .../knowledge_distillation/run_distill.sh | 1 - .../multi_task_learning/run_multi_task.sh | 1 - .../MRQA2019-D-NET/server/xlnet_server/serve.sh | 1 - .../auto_dialogue_evaluation/.run_ce.sh | 1 - .../auto_dialogue_evaluation/README.md | 4 ---- .../auto_dialogue_evaluation/run.sh | 1 - .../dialogue_general_understanding/.run_ce.sh | 6 ------ .../dialogue_general_understanding/README.md | 2 -- .../dialogue_general_understanding/run.sh | 1 - .../language_representations_kit/BERT/README.md | 3 --- .../ELMo/LAC_demo/run.sh | 2 -- PaddleNLP/lexical_analysis/.run_ce.sh | 2 -- PaddleNLP/lexical_analysis/run.sh | 2 -- PaddleNLP/lexical_analysis/run_ernie.sh | 2 -- .../transformer/README.md | 8 -------- .../transformer/train.py | 2 -- .../deep_attention_matching_net/.run_ce.sh | 1 - .../deep_attention_matching_net/douban/train.sh | 1 - .../deep_attention_matching_net/ubuntu/train.sh | 1 - .../transformer/README_cn.md | 2 -- PaddleSlim/.run_ce.sh | 1 - PaddleSlim/classification/distillation/run.sh | 4 ---- PaddleSlim/classification/quantization/run.sh | 4 ---- PaddleSlim/light_nas/run.sh | 3 --- PaddleSlim/run.sh | 4 ---- PaddleSlim/ssd/train.py | 14 -------------- 84 files changed, 1 insertion(+), 254 deletions(-) diff --git a/PaddleCV/PaddleDetection/tools/eval.py b/PaddleCV/PaddleDetection/tools/eval.py index 4c941863..6e150448 100644 --- a/PaddleCV/PaddleDetection/tools/eval.py +++ b/PaddleCV/PaddleDetection/tools/eval.py @@ -18,19 +18,6 @@ from __future__ import print_function import os - -def set_paddle_flags(**kwargs): - for key, value in kwargs.items(): - if os.environ.get(key, None) is None: - os.environ[key] = str(value) - - -# NOTE(paddle-dev): All of these flags should be set before -# `import paddle`. Otherwise, it would not take any effect. -set_paddle_flags( - FLAGS_eager_delete_tensor_gb=0, # enable GC to save memory -) - import paddle.fluid as fluid from ppdet.utils.eval_utils import parse_fetches, eval_run, eval_results, json_eval_results diff --git a/PaddleCV/PaddleDetection/tools/infer.py b/PaddleCV/PaddleDetection/tools/infer.py index 64049e3f..8097a9ca 100644 --- a/PaddleCV/PaddleDetection/tools/infer.py +++ b/PaddleCV/PaddleDetection/tools/infer.py @@ -22,19 +22,6 @@ import glob import numpy as np from PIL import Image - -def set_paddle_flags(**kwargs): - for key, value in kwargs.items(): - if os.environ.get(key, None) is None: - os.environ[key] = str(value) - - -# NOTE(paddle-dev): All of these flags should be set before -# `import paddle`. Otherwise, it would not take any effect. -set_paddle_flags( - FLAGS_eager_delete_tensor_gb=0, # enable GC to save memory -) - from paddle import fluid from ppdet.utils.cli import print_total_cfg diff --git a/PaddleCV/PaddleDetection/tools/train.py b/PaddleCV/PaddleDetection/tools/train.py index b9099210..a96d50f4 100644 --- a/PaddleCV/PaddleDetection/tools/train.py +++ b/PaddleCV/PaddleDetection/tools/train.py @@ -22,19 +22,6 @@ import numpy as np import datetime from collections import deque - -def set_paddle_flags(**kwargs): - for key, value in kwargs.items(): - if os.environ.get(key, None) is None: - os.environ[key] = str(value) - - -# NOTE(paddle-dev): All of these flags should be set before -# `import paddle`. Otherwise, it would not take any effect. -set_paddle_flags( - FLAGS_eager_delete_tensor_gb=0, # enable GC to save memory -) - from paddle import fluid from ppdet.experimental import mixed_precision_context diff --git a/PaddleCV/PaddleGAN/cycle_gan/train.py b/PaddleCV/PaddleGAN/cycle_gan/train.py index a85da0ae..5b2b4e40 100644 --- a/PaddleCV/PaddleGAN/cycle_gan/train.py +++ b/PaddleCV/PaddleGAN/cycle_gan/train.py @@ -35,13 +35,7 @@ else: # not take any effect. set_paddle_flags({ 'FLAGS_cudnn_exhaustive_search': use_cudnn_exhaustive_search, - 'FLAGS_conv_workspace_size_limit': 256, - 'FLAGS_eager_delete_tensor_gb': 0, # enable gc - # You can omit the following settings, because the default - # value of FLAGS_memory_fraction_of_eager_deletion is 1, - # and default value of FLAGS_fast_eager_deletion_mode is 1 - 'FLAGS_memory_fraction_of_eager_deletion': 1, - 'FLAGS_fast_eager_deletion_mode': 1 + 'FLAGS_conv_workspace_size_limit': 256 }) import random diff --git a/PaddleCV/PaddleGAN/scripts/run_SPADE.sh b/PaddleCV/PaddleGAN/scripts/run_SPADE.sh index 3833dc7d..312f48fd 100644 --- a/PaddleCV/PaddleGAN/scripts/run_SPADE.sh +++ b/PaddleCV/PaddleGAN/scripts/run_SPADE.sh @@ -1,4 +1,2 @@ -export FLAGS_eager_delete_tensor_gb=0.0 -export FLAGS_fast_eager_deletion_mode=1 export FLAGS_fraction_of_gpu_memory_to_use=0.01 CUDA_VISIBLE_DEVICES=0 python train.py --model_net SPADE --dataset cityscapes --train_list train_list --test_list val_list --crop_type Random --batch_size 1 --epoch 200 --load_height 612 --load_width 1124 --crop_height 512 --crop_width 1024 --label_nc 36 diff --git a/PaddleCV/PaddleVideo/models/bmn/README.md b/PaddleCV/PaddleVideo/models/bmn/README.md index 36f0857d..c146e650 100644 --- a/PaddleCV/PaddleVideo/models/bmn/README.md +++ b/PaddleCV/PaddleVideo/models/bmn/README.md @@ -29,9 +29,7 @@ BMN的训练数据采用ActivityNet1.3提供的数据集，数据下载及准备 数据准备完毕后，可以通过如下两种方式启动训练： export CUDA_VISIBLE_DEVICES=0,1,2,3 - export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 - export FLAGS_fast_eager_deletion_mode=1 python train.py --model_name=BMN \ --config=./configs/bmn.yaml \ --log_interval=10 \ diff --git a/PaddleCV/PaddleVideo/models/bsn/README.md b/PaddleCV/PaddleVideo/models/bsn/README.md index a4d0374b..7e9572d4 100644 --- a/PaddleCV/PaddleVideo/models/bsn/README.md +++ b/PaddleCV/PaddleVideo/models/bsn/README.md @@ -30,9 +30,7 @@ TEM模块以snippet-level的特征序列作为输入，预测每一个时序位 数据准备完毕后，可以通过如下两种方式启动训练： export CUDA_VISIBLE_DEVICES=0 - export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 - export FLAGS_fast_eager_deletion_mode=1 python train.py --model_name=BsnTem \ --config=./configs/bsn_tem.yaml \ --log_interval=10 \ @@ -60,9 +58,7 @@ PEM模块以PGM模块输出的BSP特征作为输入，输出proposal包含动作 数据准备完毕后，可以通过如下两种方式启动训练： export CUDA_VISIBLE_DEVICES=0 - export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 - export FLAGS_fast_eager_deletion_mode=1 python train.py --model_name=BsnPem \ --config=./configs/bsn_pem.yaml \ --log_interval=10 \ diff --git a/PaddleCV/PaddleVideo/models/ctcn/README.md b/PaddleCV/PaddleVideo/models/ctcn/README.md index bded2862..b33e902b 100644 --- a/PaddleCV/PaddleVideo/models/ctcn/README.md +++ b/PaddleCV/PaddleVideo/models/ctcn/README.md @@ -25,8 +25,6 @@ C-TCN的训练数据采用ActivityNet1.3提供的数据集，数据下载及准 数据准备完毕后，可以通过如下两种方式启动训练： export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 - export FLAGS_fast_eager_deletion_mode=1 - export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 python train.py --model_name=CTCN \ --config=./configs/ctcn.yaml \ diff --git a/PaddleCV/PaddleVideo/models/tsm/README.md b/PaddleCV/PaddleVideo/models/tsm/README.md index 0364799e..3526b9bd 100644 --- a/PaddleCV/PaddleVideo/models/tsm/README.md +++ b/PaddleCV/PaddleVideo/models/tsm/README.md @@ -35,8 +35,6 @@ TSM的训练数据采用由DeepMind公布的Kinetics-400动作识别数据集。 数据准备完毕后，可以通过如下两种方式启动训练： export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 - export FLAGS_fast_eager_deletion_mode=1 - export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 python train.py --model_name=TSM \ --config=./configs/tsm.yaml \ diff --git a/PaddleCV/PaddleVideo/models/tsn/README.md b/PaddleCV/PaddleVideo/models/tsn/README.md index 80ca3268..52bc4ff0 100644 --- a/PaddleCV/PaddleVideo/models/tsn/README.md +++ b/PaddleCV/PaddleVideo/models/tsn/README.md @@ -26,8 +26,6 @@ TSN的训练数据采用由DeepMind公布的Kinetics-400动作识别数据集。 数据准备完毕后，可以通过如下两种方式启动训练： export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 - export FLAGS_fast_eager_deletion_mode=1 - export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 python train.py --model_name=TSN \ --config=./configs/tsn.yaml \ diff --git a/PaddleCV/PaddleVideo/run.sh b/PaddleCV/PaddleVideo/run.sh index 635e3a38..66698277 100644 --- a/PaddleCV/PaddleVideo/run.sh +++ b/PaddleCV/PaddleVideo/run.sh @@ -25,8 +25,6 @@ weights="" #set the path of weights to enable eval and predicut, just ignore thi export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 #export CUDA_VISIBLE_DEVICES=0,1,2,3 #export CUDA_VISIBLE_DEVICES=0 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 if [ "$mode"x == "train"x ]; then diff --git a/PaddleCV/PaddleVideo/run_ce.sh b/PaddleCV/PaddleVideo/run_ce.sh index e848755d..548562e9 100755 --- a/PaddleCV/PaddleVideo/run_ce.sh +++ b/PaddleCV/PaddleVideo/run_ce.sh @@ -1,7 +1,5 @@ #!/bin/bash -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 export CUDA_VISIBLE_DEVICES=0 diff --git a/PaddleCV/deeplabv3+/train.py b/PaddleCV/deeplabv3+/train.py index e0a1f10b..be46915f 100755 --- a/PaddleCV/deeplabv3+/train.py +++ b/PaddleCV/deeplabv3+/train.py @@ -27,12 +27,6 @@ def set_paddle_flags(flags): # set before `import paddle`. Otherwise, it would # not take any effect. set_paddle_flags({ - 'FLAGS_eager_delete_tensor_gb': 0, # enable GC - # You can omit the following settings, because the default - # value of FLAGS_memory_fraction_of_eager_deletion is 1, - # and default value of FLAGS_fast_eager_deletion_mode is 1 - 'FLAGS_memory_fraction_of_eager_deletion': 1, - 'FLAGS_fast_eager_deletion_mode': 1, # Setting the default used gpu memory 'FLAGS_fraction_of_gpu_memory_to_use': 0.98 }) diff --git a/PaddleCV/face_detection/train.py b/PaddleCV/face_detection/train.py index c74cc8c6..4eecabe0 100644 --- a/PaddleCV/face_detection/train.py +++ b/PaddleCV/face_detection/train.py @@ -21,21 +21,6 @@ import numpy as np import time import argparse import functools - - -def set_paddle_flags(**kwargs): - for key, value in kwargs.items(): - if os.environ.get(key, None) is None: - os.environ[key] = str(value) - - -# NOTE(paddle-dev): All of these flags should be -# set before `import paddle`. Otherwise, it would -# not take any effect. -set_paddle_flags( - FLAGS_eager_delete_tensor_gb=0, # enable GC to save memory -) - import paddle import paddle.fluid as fluid from pyramidbox import PyramidBox diff --git a/PaddleCV/image_classification/scripts/train/AlexNet.sh b/PaddleCV/image_classification/scripts/train/AlexNet.sh index 6919f2b9..ac950db6 100644 --- a/PaddleCV/image_classification/scripts/train/AlexNet.sh +++ b/PaddleCV/image_classification/scripts/train/AlexNet.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® P40 8cards 120epochs 55h export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #AlexNet: diff --git a/PaddleCV/image_classification/scripts/train/EfficientNetB0.sh b/PaddleCV/image_classification/scripts/train/EfficientNetB0.sh index 37a5d9a4..18dce3ac 100644 --- a/PaddleCV/image_classification/scripts/train/EfficientNetB0.sh +++ b/PaddleCV/image_classification/scripts/train/EfficientNetB0.sh @@ -1,6 +1,4 @@ export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.96 diff --git a/PaddleCV/image_classification/scripts/train/GoogLeNet.sh b/PaddleCV/image_classification/scripts/train/GoogLeNet.sh index 63171b31..d7c8c445 100644 --- a/PaddleCV/image_classification/scripts/train/GoogLeNet.sh +++ b/PaddleCV/image_classification/scripts/train/GoogLeNet.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® V100 4cards 200epochs 132h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #GoogLeNet: diff --git a/PaddleCV/image_classification/scripts/train/InceptionV4.sh b/PaddleCV/image_classification/scripts/train/InceptionV4.sh index ba3c4954..17bf5ed9 100644 --- a/PaddleCV/image_classification/scripts/train/InceptionV4.sh +++ b/PaddleCV/image_classification/scripts/train/InceptionV4.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® V100 8cards 200epochs 367h export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #InceptionV4 diff --git a/PaddleCV/image_classification/scripts/train/MobileNetV1.sh b/PaddleCV/image_classification/scripts/train/MobileNetV1.sh index 8d00ce7c..02167a85 100644 --- a/PaddleCV/image_classification/scripts/train/MobileNetV1.sh +++ b/PaddleCV/image_classification/scripts/train/MobileNetV1.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® V100 4cards 120epochs 55h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 diff --git a/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_25.sh b/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_25.sh index aa7f74ba..d3e61d65 100644 --- a/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_25.sh +++ b/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_25.sh @@ -1,6 +1,4 @@ export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 diff --git a/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_5.sh b/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_5.sh index 85fdbfdc..56f8418b 100644 --- a/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_5.sh +++ b/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_5.sh @@ -1,6 +1,4 @@ export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 diff --git a/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_75.sh b/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_75.sh index ceeba744..14a36158 100644 --- a/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_75.sh +++ b/PaddleCV/image_classification/scripts/train/MobileNetV1_x0_75.sh @@ -1,6 +1,4 @@ export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 diff --git a/PaddleCV/image_classification/scripts/train/MobileNetV2.sh b/PaddleCV/image_classification/scripts/train/MobileNetV2.sh index 7a0ce41c..cd116ac7 100644 --- a/PaddleCV/image_classification/scripts/train/MobileNetV2.sh +++ b/PaddleCV/image_classification/scripts/train/MobileNetV2.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® V100 4cards 240epochs 135h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 diff --git a/PaddleCV/image_classification/scripts/train/MobileNetV2_x0_75.sh b/PaddleCV/image_classification/scripts/train/MobileNetV2_x0_75.sh index 511cfa71..092adf10 100644 --- a/PaddleCV/image_classification/scripts/train/MobileNetV2_x0_75.sh +++ b/PaddleCV/image_classification/scripts/train/MobileNetV2_x0_75.sh @@ -1,6 +1,4 @@ export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 diff --git a/PaddleCV/image_classification/scripts/train/ResNeXt101_32x4d.sh b/PaddleCV/image_classification/scripts/train/ResNeXt101_32x4d.sh index 91d8b5bb..192157c9 100644 --- a/PaddleCV/image_classification/scripts/train/ResNeXt101_32x4d.sh +++ b/PaddleCV/image_classification/scripts/train/ResNeXt101_32x4d.sh @@ -1,8 +1,6 @@ #Training details #Missed export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNeXt101_32x4d diff --git a/PaddleCV/image_classification/scripts/train/ResNeXt101_64x4d.sh b/PaddleCV/image_classification/scripts/train/ResNeXt101_64x4d.sh index f5aeb3a3..0c35add3 100644 --- a/PaddleCV/image_classification/scripts/train/ResNeXt101_64x4d.sh +++ b/PaddleCV/image_classification/scripts/train/ResNeXt101_64x4d.sh @@ -1,8 +1,6 @@ #Training details #Missed export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNeXt101_64x4d diff --git a/PaddleCV/image_classification/scripts/train/ResNeXt101_vd_32x4d.sh b/PaddleCV/image_classification/scripts/train/ResNeXt101_vd_32x4d.sh index 5e934480..6db60b84 100644 --- a/PaddleCV/image_classification/scripts/train/ResNeXt101_vd_32x4d.sh +++ b/PaddleCV/image_classification/scripts/train/ResNeXt101_vd_32x4d.sh @@ -1,8 +1,6 @@ #Training details #Missed export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNeXt101_vd_32x4d diff --git a/PaddleCV/image_classification/scripts/train/ResNeXt101_vd_64x4d.sh b/PaddleCV/image_classification/scripts/train/ResNeXt101_vd_64x4d.sh index f3d11779..eff71951 100644 --- a/PaddleCV/image_classification/scripts/train/ResNeXt101_vd_64x4d.sh +++ b/PaddleCV/image_classification/scripts/train/ResNeXt101_vd_64x4d.sh @@ -1,8 +1,6 @@ #Training details #Missed export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNeXt101_vd_64x4d diff --git a/PaddleCV/image_classification/scripts/train/ResNeXt152_64x4d.sh b/PaddleCV/image_classification/scripts/train/ResNeXt152_64x4d.sh index 0a1bd518..39e8a39e 100644 --- a/PaddleCV/image_classification/scripts/train/ResNeXt152_64x4d.sh +++ b/PaddleCV/image_classification/scripts/train/ResNeXt152_64x4d.sh @@ -1,7 +1,5 @@ #Training details export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNeXt152_64x4d diff --git a/PaddleCV/image_classification/scripts/train/ResNet101.sh b/PaddleCV/image_classification/scripts/train/ResNet101.sh index a2af4385..28b8ccfa 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet101.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet101.sh @@ -1,8 +1,6 @@ ##Training details #GPU: NVIDIA® Tesla® V100 4cards 120epochs 100h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNet101: diff --git a/PaddleCV/image_classification/scripts/train/ResNet101_vd.sh b/PaddleCV/image_classification/scripts/train/ResNet101_vd.sh index b9bdf778..a7dc7983 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet101_vd.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet101_vd.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® V100 4cards 200epochs 182h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNet101_vd diff --git a/PaddleCV/image_classification/scripts/train/ResNet152.sh b/PaddleCV/image_classification/scripts/train/ResNet152.sh index 44275753..373cecb8 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet152.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet152.sh @@ -1,8 +1,6 @@ ##Training details #GPU: NVIDIA® Tesla® P40 8cards 120epochs 200h export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNet152: diff --git a/PaddleCV/image_classification/scripts/train/ResNet152_vd.sh b/PaddleCV/image_classification/scripts/train/ResNet152_vd.sh index b4cb84ad..500b4628 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet152_vd.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet152_vd.sh @@ -1,8 +1,6 @@ ##Training details #GPU: NVIDIA® Tesla® P40 8cards 200epochs 346h export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 python train.py \ diff --git a/PaddleCV/image_classification/scripts/train/ResNet18.sh b/PaddleCV/image_classification/scripts/train/ResNet18.sh index b3d1018c..e6cfe9a7 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet18.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet18.sh @@ -1,8 +1,6 @@ ##Training details #GPU: NVIDIA® Tesla® V100 4cards 120epochs 67h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNet18: diff --git a/PaddleCV/image_classification/scripts/train/ResNet18_vd.sh b/PaddleCV/image_classification/scripts/train/ResNet18_vd.sh index c95b9325..4bc122d2 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet18_vd.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet18_vd.sh @@ -1,6 +1,4 @@ export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 python train.py \ diff --git a/PaddleCV/image_classification/scripts/train/ResNet200_vd.sh b/PaddleCV/image_classification/scripts/train/ResNet200_vd.sh index 464db8ac..cb9f14da 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet200_vd.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet200_vd.sh @@ -1,8 +1,6 @@ #Training details #Machine: Missed export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNet200_vd diff --git a/PaddleCV/image_classification/scripts/train/ResNet34.sh b/PaddleCV/image_classification/scripts/train/ResNet34.sh index 5ce4689b..f4f2b0d6 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet34.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet34.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® V100 4cards 120epochs 73h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNet34: python train.py \ diff --git a/PaddleCV/image_classification/scripts/train/ResNet34_vd.sh b/PaddleCV/image_classification/scripts/train/ResNet34_vd.sh index 56a31b69..14df38da 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet34_vd.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet34_vd.sh @@ -1,6 +1,4 @@ export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 python train.py \ diff --git a/PaddleCV/image_classification/scripts/train/ResNet50.sh b/PaddleCV/image_classification/scripts/train/ResNet50.sh index 47063075..1c68daec 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet50.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet50.sh @@ -1,8 +1,6 @@ ##Training details #GPU: NVIDIA® Tesla® V100 4cards 120epochs 67h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNet50: diff --git a/PaddleCV/image_classification/scripts/train/ResNet50_vc.sh b/PaddleCV/image_classification/scripts/train/ResNet50_vc.sh index d5d0cc5e..b12c0dbe 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet50_vc.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet50_vc.sh @@ -1,8 +1,6 @@ ##Training details #GPU: NVIDIA® Tesla® V100 4cards 200epochs 141h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #ResNet50_vc diff --git a/PaddleCV/image_classification/scripts/train/ResNet50_vd.sh b/PaddleCV/image_classification/scripts/train/ResNet50_vd.sh index 968e3dd0..83df7473 100644 --- a/PaddleCV/image_classification/scripts/train/ResNet50_vd.sh +++ b/PaddleCV/image_classification/scripts/train/ResNet50_vd.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® V100 4cards 200epochs 120h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 python train.py \ diff --git a/PaddleCV/image_classification/scripts/train/SENet154_vd.sh b/PaddleCV/image_classification/scripts/train/SENet154_vd.sh index a363a108..7ad2b502 100644 --- a/PaddleCV/image_classification/scripts/train/SENet154_vd.sh +++ b/PaddleCV/image_classification/scripts/train/SENet154_vd.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® P40 8cards 200epochs 916h export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #SE_154 diff --git a/PaddleCV/image_classification/scripts/train/SE_ResNeXt101_32x4d.sh b/PaddleCV/image_classification/scripts/train/SE_ResNeXt101_32x4d.sh index a385814a..38701120 100644 --- a/PaddleCV/image_classification/scripts/train/SE_ResNeXt101_32x4d.sh +++ b/PaddleCV/image_classification/scripts/train/SE_ResNeXt101_32x4d.sh @@ -1,8 +1,6 @@ ##Training details #GPU: NVIDIA® Tesla® P40 8cards 120epochs 566h export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #SE_ResNeXt101_32x4d: diff --git a/PaddleCV/image_classification/scripts/train/SE_ResNeXt50_32x4d.sh b/PaddleCV/image_classification/scripts/train/SE_ResNeXt50_32x4d.sh index acfadb80..f0d3b61b 100644 --- a/PaddleCV/image_classification/scripts/train/SE_ResNeXt50_32x4d.sh +++ b/PaddleCV/image_classification/scripts/train/SE_ResNeXt50_32x4d.sh @@ -1,8 +1,6 @@ #Training details #Machine:V100 4cards 200epochs 282h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 diff --git a/PaddleCV/image_classification/scripts/train/ShuffleNetV2.sh b/PaddleCV/image_classification/scripts/train/ShuffleNetV2.sh index 369e5879..91a95ab2 100644 --- a/PaddleCV/image_classification/scripts/train/ShuffleNetV2.sh +++ b/PaddleCV/image_classification/scripts/train/ShuffleNetV2.sh @@ -1,8 +1,6 @@ ##Training details #GPU: NVIDIA® Tesla® K40 4cards 240epochs 156h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 python train.py \ diff --git a/PaddleCV/image_classification/scripts/train/ShuffleNetV2_swish.sh b/PaddleCV/image_classification/scripts/train/ShuffleNetV2_swish.sh index b3e29dd3..c2f13db1 100644 --- a/PaddleCV/image_classification/scripts/train/ShuffleNetV2_swish.sh +++ b/PaddleCV/image_classification/scripts/train/ShuffleNetV2_swish.sh @@ -1,8 +1,6 @@ ##Training details #GPU: NVIDIA® Tesla® K40 4cards 240epochs 156h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 python train.py \ diff --git a/PaddleCV/image_classification/scripts/train/VGG11.sh b/PaddleCV/image_classification/scripts/train/VGG11.sh index ad8934e4..f4b79a3d 100644 --- a/PaddleCV/image_classification/scripts/train/VGG11.sh +++ b/PaddleCV/image_classification/scripts/train/VGG11.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® P40 8cards 90epochs 52h export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #VGG11: diff --git a/PaddleCV/image_classification/scripts/train/VGG13.sh b/PaddleCV/image_classification/scripts/train/VGG13.sh index 24960f88..17784de7 100644 --- a/PaddleCV/image_classification/scripts/train/VGG13.sh +++ b/PaddleCV/image_classification/scripts/train/VGG13.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® V100 4cards 90epochs 58h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #VGG13: diff --git a/PaddleCV/image_classification/scripts/train/VGG16.sh b/PaddleCV/image_classification/scripts/train/VGG16.sh index ebf5a356..4f12a7e4 100644 --- a/PaddleCV/image_classification/scripts/train/VGG16.sh +++ b/PaddleCV/image_classification/scripts/train/VGG16.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® P40 8cards 90epochs 72h export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #VGG16: diff --git a/PaddleCV/image_classification/scripts/train/VGG19.sh b/PaddleCV/image_classification/scripts/train/VGG19.sh index bca6a002..6bbacc1f 100644 --- a/PaddleCV/image_classification/scripts/train/VGG19.sh +++ b/PaddleCV/image_classification/scripts/train/VGG19.sh @@ -1,8 +1,6 @@ #Training details #GPU: NVIDIA® Tesla® V100 4cards 150epochs 173h export CUDA_VISIBLE_DEVICES=0,1,2,3 -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export FLAGS_fraction_of_gpu_memory_to_use=0.98 #VGG19: diff --git a/PaddleCV/image_classification/train.py b/PaddleCV/image_classification/train.py index dfd0f591..ec002ff2 100755 --- a/PaddleCV/image_classification/train.py +++ b/PaddleCV/image_classification/train.py @@ -31,7 +31,6 @@ def set_paddle_flags(flags): # set before `import paddle`. Otherwise, it would # not take any effect. set_paddle_flags({ - 'FLAGS_eager_delete_tensor_gb': 0, # enable gc 'FLAGS_fraction_of_gpu_memory_to_use': 0.98 }) diff --git a/PaddleCV/rcnn/train.py b/PaddleCV/rcnn/train.py index 705ad33a..f2d62062 100644 --- a/PaddleCV/rcnn/train.py +++ b/PaddleCV/rcnn/train.py @@ -26,8 +26,6 @@ def set_paddle_flags(flags): set_paddle_flags({ 'FLAGS_conv_workspace_size_limit': 500, - 'FLAGS_eager_delete_tensor_gb': 0, # enable gc - 'FLAGS_memory_fraction_of_eager_deletion': 1, 'FLAGS_fraction_of_gpu_memory_to_use': 0.98 }) diff --git a/PaddleCV/ssd/train.py b/PaddleCV/ssd/train.py index 7bbb0fda..0e7c7958 100644 --- a/PaddleCV/ssd/train.py +++ b/PaddleCV/ssd/train.py @@ -20,20 +20,6 @@ import shutil import math import multiprocessing - -def set_paddle_flags(**kwargs): - for key, value in kwargs.items(): - if os.environ.get(key, None) is None: - os.environ[key] = str(value) - - -# NOTE(paddle-dev): All of these flags should be -# set before `import paddle`. Otherwise, it would -# not take any effect. -set_paddle_flags( - FLAGS_eager_delete_tensor_gb=0, # enable GC to save memory -) - import paddle import paddle.fluid as fluid import reader diff --git a/PaddleCV/yolov3/train.py b/PaddleCV/yolov3/train.py index 5f2284cf..9bb9c062 100644 --- a/PaddleCV/yolov3/train.py +++ b/PaddleCV/yolov3/train.py @@ -25,8 +25,6 @@ def set_paddle_flags(flags): set_paddle_flags({ - 'FLAGS_eager_delete_tensor_gb': 0, # enable gc - 'FLAGS_memory_fraction_of_eager_deletion': 1, 'FLAGS_fraction_of_gpu_memory_to_use': 0.98 }) diff --git a/PaddleNLP/Research/ACL2018-DAM/run.sh b/PaddleNLP/Research/ACL2018-DAM/run.sh index 47395cb7..4c42b047 100755 --- a/PaddleNLP/Research/ACL2018-DAM/run.sh +++ b/PaddleNLP/Research/ACL2018-DAM/run.sh @@ -1,5 +1,4 @@ export CUDA_VISIBLE_DEVICES=3 -export FLAGS_eager_delete_tensor_gb=0.0 #train on ubuntu python -u main.py \ diff --git a/PaddleNLP/Research/ACL2018-DAM/run_CPU.sh b/PaddleNLP/Research/ACL2018-DAM/run_CPU.sh index 091eda1f..220ec6a1 100755 --- a/PaddleNLP/Research/ACL2018-DAM/run_CPU.sh +++ b/PaddleNLP/Research/ACL2018-DAM/run_CPU.sh @@ -1,5 +1,4 @@ export CPU_NUM=1 -export FLAGS_eager_delete_tensor_gb=0.0 #train on ubuntu python -u main.py \ diff --git a/PaddleNLP/Research/MRQA2019-BASELINE/run_finetuning.sh b/PaddleNLP/Research/MRQA2019-BASELINE/run_finetuning.sh index bbe9bfe1..71e191f5 100644 --- a/PaddleNLP/Research/MRQA2019-BASELINE/run_finetuning.sh +++ b/PaddleNLP/Research/MRQA2019-BASELINE/run_finetuning.sh @@ -18,7 +18,6 @@ set -xe export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 # set CUDA_VISIBLE_DEVICES export CUDA_VISIBLE_DEVICES=0 diff --git a/PaddleNLP/Research/MRQA2019-BASELINE/run_predict.sh b/PaddleNLP/Research/MRQA2019-BASELINE/run_predict.sh index df5e86b5..6c11820e 100644 --- a/PaddleNLP/Research/MRQA2019-BASELINE/run_predict.sh +++ b/PaddleNLP/Research/MRQA2019-BASELINE/run_predict.sh @@ -18,7 +18,6 @@ set -xe export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 # set CUDA_VISIBLE_DEVICES export CUDA_VISIBLE_DEVICES=0 diff --git a/PaddleNLP/Research/MRQA2019-D-NET/knowledge_distillation/run_distill.sh b/PaddleNLP/Research/MRQA2019-D-NET/knowledge_distillation/run_distill.sh index 9042506c..8b040392 100755 --- a/PaddleNLP/Research/MRQA2019-D-NET/knowledge_distillation/run_distill.sh +++ b/PaddleNLP/Research/MRQA2019-D-NET/knowledge_distillation/run_distill.sh @@ -1,7 +1,6 @@ #!/bin/bash export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 if [ ! "$CUDA_VISIBLE_DEVICES" ] diff --git a/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/run_multi_task.sh b/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/run_multi_task.sh index e65d4cc2..fe9d9023 100755 --- a/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/run_multi_task.sh +++ b/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/run_multi_task.sh @@ -2,7 +2,6 @@ # for gpu memory optimization export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 diff --git a/PaddleNLP/Research/MRQA2019-D-NET/server/xlnet_server/serve.sh b/PaddleNLP/Research/MRQA2019-D-NET/server/xlnet_server/serve.sh index ff56192b..f9beecea 100755 --- a/PaddleNLP/Research/MRQA2019-D-NET/server/xlnet_server/serve.sh +++ b/PaddleNLP/Research/MRQA2019-D-NET/server/xlnet_server/serve.sh @@ -1,5 +1,4 @@ export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 export FLAGS_fraction_of_gpu_memory_to_use=0.1 python serve.py ./infer_model_800_bs128 5001 & diff --git a/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/.run_ce.sh b/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/.run_ce.sh index fea71c7a..5b4881ce 100644 --- a/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/.run_ce.sh +++ b/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/.run_ce.sh @@ -1,7 +1,6 @@ #!/bin/bash export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1.0 export CUDA_VISIBLE_DEVICES=0 diff --git a/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/README.md b/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/README.md index 88e4eb64..bd3fbcd4 100644 --- a/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/README.md +++ b/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/README.md @@ -156,7 +156,6 @@ export CUDA_VISIBLE_DEVICES=0,1,2,3 ``` export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 #开启显存优化 export CUDA_VISIBLE_DEVICES=0 #GPU单卡训练 #export CUDA_VISIBLE_DEVICES=0,1,2,3 #GPU多卡训练 @@ -222,7 +221,6 @@ task_type: train、predict、evaluate、inference, 选择4个参数选项中任 ``` export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 #开启显存优化 export CUDA_VISIBLE_DEVICES=0 #GPU单卡训练 #export CUDA_VISIBLE_DEVICES=0,1,2,3 #GPU多卡训练 @@ -298,7 +296,6 @@ export CUDA_VISIBLE_DEVICES=0 #用户可自行指定空闲的卡 ``` export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 #开启显存优化 export CUDA_VISIBLE_DEVICES=0 #单卡预测 #export CUDA_VISIBLE_DEVICES= #CPU预测 @@ -346,7 +343,6 @@ task_type: train、predict、evaluate、inference, 选择4个参数选项中任 ``` export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 #开启显存优化 export CUDA_VISIBLE_DEVICES=0 #单卡预测 #export CUDA_VISIBLE_DEVICES= #CPU预测 diff --git a/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/run.sh b/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/run.sh index 78f8da40..52c30b7b 100755 --- a/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/run.sh +++ b/PaddleNLP/dialogue_model_toolkit/auto_dialogue_evaluation/run.sh @@ -1,7 +1,6 @@ #!/bin/bash export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1.0 export CUDA_VISIBLE_DEVICES=0 diff --git a/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/.run_ce.sh b/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/.run_ce.sh index 8ce98ff7..4feb9560 100644 --- a/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/.run_ce.sh +++ b/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/.run_ce.sh @@ -52,12 +52,6 @@ train_mrda(){ --enable_ce=store_true } -# FIXME(zjl): this model would fail when GC is enabled, -# but it seems that this error is from the model itself. -# See issue here: https://github.com/PaddlePaddle/Paddle/issues/18994#event-2532039900 -# To fix ce, disable gc in this model temporarily. -export FLAGS_eager_delete_tensor_gb=1 - cudaid=${multi:=0,1,2,3} export CUDA_VISIBLE_DEVICES=$cudaid train_atis_slot | python _ce.py diff --git a/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/README.md b/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/README.md index 77783058..68f4b8e7 100644 --- a/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/README.md +++ b/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/README.md @@ -177,7 +177,6 @@ export CUDA_VISIBLE_DEVICES=0,1,2,3 ``` export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 #开启显存优化 export CUDA_VISIBLE_DEVICES=0 #GPU单卡训练 #export CUDA_VISIBLE_DEVICES=0,1,2,3 #GPU多卡训练 @@ -259,7 +258,6 @@ export CUDA_VISIBLE_DEVICES=0 ``` export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 #开启显存优化 export CUDA_VISIBLE_DEVICES=0 #单卡预测 #export CUDA_VISIBLE_DEVICES= #CPU预测 diff --git a/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/run.sh b/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/run.sh index 1cf3aa3f..82467477 100644 --- a/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/run.sh +++ b/PaddleNLP/dialogue_model_toolkit/dialogue_general_understanding/run.sh @@ -1,7 +1,6 @@ #!/bin/bash export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 export CUDA_VISIBLE_DEVICES=0 if [ ! "$CUDA_VISIBLE_DEVICES" ] diff --git a/PaddleNLP/language_representations_kit/BERT/README.md b/PaddleNLP/language_representations_kit/BERT/README.md index b7eb9d8c..64df0882 100644 --- a/PaddleNLP/language_representations_kit/BERT/README.md +++ b/PaddleNLP/language_representations_kit/BERT/README.md @@ -147,7 +147,6 @@ export current_endpoint=192.168.0.17:9185 ```shell export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 BERT_BASE_PATH="chinese_L-12_H-768_A-12" @@ -209,7 +208,6 @@ SQuAD v1.1 ```shell export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 export CUDA_VISIBLE_DEVICES=0,1,2,3 BERT_BASE_PATH="uncased_L-12_H-768_A-12" @@ -255,7 +253,6 @@ python ${SQUAD_PATH}/evaluate-v1.1.py ${SQUAD_PATH}/dev-v1.1.json ${CHECKPOINT_P ```shell export FLAGS_sync_nccl_allreduce=0 -export FLAGS_eager_delete_tensor_gb=1 export CUDA_VISIBLE_DEVICES=0,1,2,3 BERT_BASE_PATH="uncased_L-12_H-768_A-12" CHECKPOINT_PATH=/path/to/save/checkpoints/ diff --git a/PaddleNLP/language_representations_kit/ELMo/LAC_demo/run.sh b/PaddleNLP/language_representations_kit/ELMo/LAC_demo/run.sh index 4d71d0c6..6e3cb629 100755 --- a/PaddleNLP/language_representations_kit/ELMo/LAC_demo/run.sh +++ b/PaddleNLP/language_representations_kit/ELMo/LAC_demo/run.sh @@ -1,6 +1,4 @@ export FLAGS_fraction_of_gpu_memory_to_use=0.5 -export FLAGS_eager_delete_tensor_gb=0.0 -export FLAGS_fast_eager_deletion_mode=1 export CUDA_VISIBLE_DEVICES=0 python train.py \ diff --git a/PaddleNLP/lexical_analysis/.run_ce.sh b/PaddleNLP/lexical_analysis/.run_ce.sh index 73f0005a..cbf7a03a 100644 --- a/PaddleNLP/lexical_analysis/.run_ce.sh +++ b/PaddleNLP/lexical_analysis/.run_ce.sh @@ -1,7 +1,5 @@ #!/bin/bash export FLAGS_fraction_of_gpu_memory_to_use=0.5 -export FLAGS_eager_delete_tensor_gb=0.0 -export FLAGS_fast_eager_deletion_mode=1 train() diff --git a/PaddleNLP/lexical_analysis/run.sh b/PaddleNLP/lexical_analysis/run.sh index 614b8ce6..99bf7b63 100644 --- a/PaddleNLP/lexical_analysis/run.sh +++ b/PaddleNLP/lexical_analysis/run.sh @@ -1,7 +1,5 @@ #!/bin/bash export FLAGS_fraction_of_gpu_memory_to_use=0.02 -export FLAGS_eager_delete_tensor_gb=0.0 -export FLAGS_fast_eager_deletion_mode=1 export CUDA_VISIBLE_DEVICES=0,1,2,3 # which GPU to use function run_train() { diff --git a/PaddleNLP/lexical_analysis/run_ernie.sh b/PaddleNLP/lexical_analysis/run_ernie.sh index 61a8e8e9..2203f4d6 100644 --- a/PaddleNLP/lexical_analysis/run_ernie.sh +++ b/PaddleNLP/lexical_analysis/run_ernie.sh @@ -1,7 +1,5 @@ #set -eux export FLAGS_fraction_of_gpu_memory_to_use=0.02 -export FLAGS_eager_delete_tensor_gb=0.0 -export FLAGS_fast_eager_deletion_mode=1 # export FLAGS_sync_nccl_allreduce=1 # export NCCL_DEBUG=INFO # export NCCL_IB_GID_INDEX=3 diff --git a/PaddleNLP/neural_machine_translation/transformer/README.md b/PaddleNLP/neural_machine_translation/transformer/README.md index 27c59a97..e5413059 100644 --- a/PaddleNLP/neural_machine_translation/transformer/README.md +++ b/PaddleNLP/neural_machine_translation/transformer/README.md @@ -68,8 +68,6 @@ 以提供的英德翻译数据为例，可以执行以下命令进行模型训练： ```sh -# open garbage collection to save memory -export FLAGS_eager_delete_tensor_gb=0.0 # setting visible devices for training export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 @@ -86,8 +84,6 @@ python -u main.py \ 以上命令中传入了执行训练（`do_train`）、训练轮数（`epoch`）和训练数据文件路径（注意请正确设置，支持通配符）等参数，更多参数的使用以及支持的模型超参数可以参见 `transformer.yaml` 配置文件，其中默认提供了 Transformer base model 的配置，如需调整可以在配置文件中更改或通过命令行传入（命令行传入内容将覆盖配置文件中的设置）。可以通过以下命令来训练 Transformer 论文中的 big model： ```sh -# open garbage collection to save memory -export FLAGS_eager_delete_tensor_gb=0.0 # setting visible devices for training export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 @@ -124,8 +120,6 @@ python -u main.py \ 以英德翻译数据为例，模型训练完成后可以执行以下命令对指定文件中的文本进行翻译： ```sh -# open garbage collection to save memory -export FLAGS_eager_delete_tensor_gb=0.0 # setting visible devices for prediction export CUDA_VISIBLE_DEVICES=0 @@ -145,8 +139,6 @@ python -u main.py \ 由 `predict_file` 指定的文件中文本的翻译结果会输出到 `output_file` 指定的文件。执行预测时需要设置 `init_from_params` 来给出模型所在目录，更多参数的使用可以在 `transformer.yaml` 文件中查阅注释说明并进行更改设置。注意若在执行预测时设置了模型超参数，应与模型训练时的设置一致，如若训练时使用 big model 的参数设置，则预测时对应类似如下命令： ```sh -# open garbage collection to save memory -export FLAGS_eager_delete_tensor_gb=0.0 # setting visible devices for prediction export CUDA_VISIBLE_DEVICES=0 diff --git a/PaddleNLP/neural_machine_translation/transformer/train.py b/PaddleNLP/neural_machine_translation/transformer/train.py index 57fa3b91..74f82504 100644 --- a/PaddleNLP/neural_machine_translation/transformer/train.py +++ b/PaddleNLP/neural_machine_translation/transformer/train.py @@ -32,8 +32,6 @@ import dist_utils import reader from transformer import create_net, position_encoding_init -if os.environ.get('FLAGS_eager_delete_tensor_gb', None) is None: - os.environ['FLAGS_eager_delete_tensor_gb'] = '0' # num_trainers is used for multi-process gpu training num_trainers = int(os.environ.get('PADDLE_TRAINERS_NUM', 1)) diff --git a/PaddleNLP/unarchived/deep_attention_matching_net/.run_ce.sh b/PaddleNLP/unarchived/deep_attention_matching_net/.run_ce.sh index 6c1c0a34..0b927bdb 100755 --- a/PaddleNLP/unarchived/deep_attention_matching_net/.run_ce.sh +++ b/PaddleNLP/unarchived/deep_attention_matching_net/.run_ce.sh @@ -3,7 +3,6 @@ export CE_MODE_X=1 export CUDA_VISIBLE_DEVICES=0 -export FLAGS_eager_delete_tensor_gb=0.0 if [ ! -e data_small.pkl ]; then wget -c http://dam-data.bj.bcebos.com/data_small.pkl fi diff --git a/PaddleNLP/unarchived/deep_attention_matching_net/douban/train.sh b/PaddleNLP/unarchived/deep_attention_matching_net/douban/train.sh index 6ed91319..eb8a7d3a 100644 --- a/PaddleNLP/unarchived/deep_attention_matching_net/douban/train.sh +++ b/PaddleNLP/unarchived/deep_attention_matching_net/douban/train.sh @@ -1,5 +1,4 @@ export CUDA_VISIBLE_DEVICES=0 -export FLAGS_eager_delete_tensor_gb=0.0 python -u ../train_and_evaluate.py --use_cuda \ --data_path ./data/data.pkl \ --ext_eval \ diff --git a/PaddleNLP/unarchived/deep_attention_matching_net/ubuntu/train.sh b/PaddleNLP/unarchived/deep_attention_matching_net/ubuntu/train.sh index 66ebc2e6..4d8b54b9 100644 --- a/PaddleNLP/unarchived/deep_attention_matching_net/ubuntu/train.sh +++ b/PaddleNLP/unarchived/deep_attention_matching_net/ubuntu/train.sh @@ -1,5 +1,4 @@ export CUDA_VISIBLE_DEVICES=0 -export FLAGS_eager_delete_tensor_gb=0.0 python -u ../train_and_evaluate.py --use_cuda \ --data_path ./data/data.pkl \ --word_emb_init ./data/word_embedding.pkl \ diff --git a/PaddleNLP/unarchived/neural_machine_translation/transformer/README_cn.md b/PaddleNLP/unarchived/neural_machine_translation/transformer/README_cn.md index bdac7cb0..8db20d95 100644 --- a/PaddleNLP/unarchived/neural_machine_translation/transformer/README_cn.md +++ b/PaddleNLP/unarchived/neural_machine_translation/transformer/README_cn.md @@ -98,8 +98,6 @@ python train.py --help ```sh # 显存使用的比例，显存不足可适当增大，最大为1 export FLAGS_fraction_of_gpu_memory_to_use=1.0 -# 显存清理的阈值，显存不足可适当减小，最小为0，为负数时不启用 -export FLAGS_eager_delete_tensor_gb=0.8 python -u train.py \ --src_vocab_fpath gen_data/wmt16_ende_data_bpe/vocab_all.bpe.32000 \ --trg_vocab_fpath gen_data/wmt16_ende_data_bpe/vocab_all.bpe.32000 \ diff --git a/PaddleSlim/.run_ce.sh b/PaddleSlim/.run_ce.sh index 3e203480..b45fbaf6 100755 --- a/PaddleSlim/.run_ce.sh +++ b/PaddleSlim/.run_ce.sh @@ -1,7 +1,6 @@ #!/bin/bash # This file is only used for continuous evaluation. -export FLAGS_eager_delete_tensor_gb=0.0 export CUDA_VISIBLE_DEVICES=3 if [ ! -d 'pretrain' ]; then diff --git a/PaddleSlim/classification/distillation/run.sh b/PaddleSlim/classification/distillation/run.sh index e9d88f50..dc899ccc 100644 --- a/PaddleSlim/classification/distillation/run.sh +++ b/PaddleSlim/classification/distillation/run.sh @@ -18,10 +18,6 @@ fi cd - -# enable GC strategy -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 - # for distillation #----------------- export CUDA_VISIBLE_DEVICES=0,1,2,3 diff --git a/PaddleSlim/classification/quantization/run.sh b/PaddleSlim/classification/quantization/run.sh index adf67f38..bbe24e62 100644 --- a/PaddleSlim/classification/quantization/run.sh +++ b/PaddleSlim/classification/quantization/run.sh @@ -30,10 +30,6 @@ fi cd - -# enable GC strategy -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 - export CUDA_VISIBLE_DEVICES=0 ## for quantization for mobilenet_v1 diff --git a/PaddleSlim/light_nas/run.sh b/PaddleSlim/light_nas/run.sh index 396247cf..5b4b4257 100644 --- a/PaddleSlim/light_nas/run.sh +++ b/PaddleSlim/light_nas/run.sh @@ -1,5 +1,2 @@ -# enable GC strategy -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 export CUDA_VISIBLE_DEVICES=0,1,2,3 python search.py diff --git a/PaddleSlim/run.sh b/PaddleSlim/run.sh index 89bffa4d..fa909323 100644 --- a/PaddleSlim/run.sh +++ b/PaddleSlim/run.sh @@ -24,10 +24,6 @@ fi cd - -# enable GC strategy -export FLAGS_fast_eager_deletion_mode=1 -export FLAGS_eager_delete_tensor_gb=0.0 - # for distillation #----------------- export CUDA_VISIBLE_DEVICES=0,1,2,3 diff --git a/PaddleSlim/ssd/train.py b/PaddleSlim/ssd/train.py index 96810ae0..c4494b5c 100644 --- a/PaddleSlim/ssd/train.py +++ b/PaddleSlim/ssd/train.py @@ -21,20 +21,6 @@ import math import multiprocessing from paddle.fluid.contrib.slim import Compressor - -def set_paddle_flags(**kwargs): - for key, value in kwargs.items(): - if os.environ.get(key, None) is None: - os.environ[key] = str(value) - - -# NOTE(paddle-dev): All of these flags should be -# set before `import paddle`. Otherwise, it would -# not take any effect. -set_paddle_flags( - FLAGS_eager_delete_tensor_gb=0, # enable GC to save memory -) - import paddle import paddle.fluid as fluid import reader -- GitLab