From 93fbd09c96f21fdca06069ab978946a570a7b65f Mon Sep 17 00:00:00 2001
From: deepr <hexiangdong2020@outlook.com>
Date: Tue, 20 Jul 2021 23:02:41 +0800
Subject: [PATCH] =?UTF-8?q?=E8=A7=A3=E5=86=B3utils=E5=BC=95=E7=94=A8?=
=?UTF-8?q?=E5=86=B2=E7=AA=81?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
---
main.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/main.py b/main.py
index f4ad80b..2a920a4 100644
--- a/main.py
+++ b/main.py
@@ -14,7 +14,7 @@ from unet_medical.unet_model import UNetMedical
from nets.deeplab_v3 import deeplab_v3
from dataset import GetDatasetGenerator
from loss import SoftmaxCrossEntropyLoss
-from utils.learning_rates import exponential_lr
+import utils.learning_rates as learning_rates
context.set_context(mode=context.PYNATIVE_MODE, save_graphs=False,
device_target='Ascend', device_id=7)
@@ -25,7 +25,7 @@ train_dataset_generator = GetDatasetGenerator('./datasets', 'train')
train_dataset = ds.GeneratorDataset(train_dataset_generator, ["data", "label"], shuffle=True)
train_dataset = train_dataset.batch(4, drop_remainder=True)
-lr_iter = exponential_lr(3e-5, 20, 0.98, 500, staircase=True)
+lr_iter = learning_rates.exponential_lr(3e-5, 20, 0.98, 500, staircase=True)
net_loss = SoftmaxCrossEntropyLoss(6, 255)
net_opt = nn.Adam(net.trainable_params(), learning_rate=lr_iter)
--
GitLab