From c11e064e18c6f50004c6d90100d5cbe0ae0dcfa6 Mon Sep 17 00:00:00 2001 From: wangmin Date: Tue, 1 Sep 2020 16:39:47 +0800 Subject: [PATCH] fix readme file for resnet_thor --- model_zoo/official/cv/resnet_thor/README.md | 16 ++++++++-------- model_zoo/official/cv/resnet_thor/train.py | 2 +- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/model_zoo/official/cv/resnet_thor/README.md b/model_zoo/official/cv/resnet_thor/README.md index d8f8d8a42b0..292382a66e2 100644 --- a/model_zoo/official/cv/resnet_thor/README.md +++ b/model_zoo/official/cv/resnet_thor/README.md @@ -58,7 +58,7 @@ After installing MindSpore via the official website, you can start training and - Running on Ascend ```python # run distributed training example -sh scripts/run_distribute_train.sh [RANK_TABLE_FILE] [DATASET_PATH] [DEVICE_NUM] +sh run_distribute_train.sh [RANK_TABLE_FILE] [DATASET_PATH] [DEVICE_NUM] # run evaluation example sh run_eval.sh [DATASET_PATH] [CHECKPOINT_PATH] @@ -68,7 +68,7 @@ sh run_eval.sh [DATASET_PATH] [CHECKPOINT_PATH] - Running on GPU ```python # run distributed training example -sh scripts/run_distribute_train_gpu.sh [DATASET_PATH] [DEVICE_NUM] +sh run_distribute_train_gpu.sh [DATASET_PATH] [DEVICE_NUM] # run evaluation example sh run_eval_gpu.sh [DATASET_PATH] [CHECKPOINT_PATH] @@ -132,18 +132,18 @@ Parameters for both training and inference can be set in config.py. "loss_scale": 128, # loss scale "momentum": 0.9, # momentum of THOR optimizer "weight_decay": 5e-4, # weight decay -"epoch_size": 45, # only valid for taining, which is always 1 for inference +"epoch_size": 40, # only valid for taining, which is always 1 for inference "save_checkpoint": True, # whether save checkpoint or not "save_checkpoint_epochs": 1, # the epoch interval between two checkpoints. By default, the checkpoint will be saved every epoch "keep_checkpoint_max": 15, # only keep the last keep_checkpoint_max checkpoint "save_checkpoint_path": "./", # path to save checkpoint relative to the executed path "label_smooth": True, # label smooth "label_smooth_factor": 0.1, # label smooth factor -"lr_init": 0.04, # learning rate init value -"lr_decay": 5, # learning rate decay rate value -"lr_end_epoch": 58, # learning rate end epoch value -"damping_init": 0.02, # damping init value for Fisher information matrix -"damping_decay": 0.87, # damping decay rate +"lr_init": 0.05672, # learning rate init value +"lr_decay": 4.9687, # learning rate decay rate value +"lr_end_epoch": 50, # learning rate end epoch value +"damping_init": 0.02345, # damping init value for Fisher information matrix +"damping_decay": 0.5467, # damping decay rate "frequency": 834, # the step interval to update second-order information matrix ``` ### Training Process diff --git a/model_zoo/official/cv/resnet_thor/train.py b/model_zoo/official/cv/resnet_thor/train.py index 025431d46c7..03d545eca14 100644 --- a/model_zoo/official/cv/resnet_thor/train.py +++ b/model_zoo/official/cv/resnet_thor/train.py @@ -118,7 +118,7 @@ if __name__ == '__main__': # define net step_size = dataset.get_dataset_size() - damping = get_model_damping(0, config.damping_init, config.damping_decay, 90, step_size) + damping = get_model_damping(0, config.damping_init, config.damping_decay, 70, step_size) lr = get_model_lr(0, config.lr_init, config.lr_decay, config.lr_end_epoch, step_size, decay_epochs=39) net = resnet50(class_num=config.class_num, damping=damping, loss_scale=config.loss_scale, frequency=config.frequency, batch_size=config.batch_size)