!14528 add ssd, resnet,unet evaluation while training process

From: @zhao_ting_v
Reviewed-by: @c_34,@wuxuejian
Signed-off-by: @c_34

model_zoo/official/cv/resnet/README.md

@@ -155,6 +155,7 @@ python eval.py --net=[resnet50|resnet101] --dataset=[cifar10|imagenet2012] --dat
   ├── src
     ├── config.py                          # parameter configuration
     ├── dataset.py                         # data preprocessing
+    ├─  eval_callback.py                   # evaluation callback while training
     ├── CrossEntropySmooth.py              # loss definition for ImageNet2012 dataset
     ├── lr_generator.py                    # generate learning rate for each step
     ├── resnet.py                          # resnet backbone, including resnet50 and resnet101 and se-resnet50
@@ -323,6 +324,10 @@ bash run_parameter_server_train.sh [resnet18|resnet50|resnet101] [cifar10|imagen
 bash run_parameter_server_train_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
 ```
 
+#### Evaluation while training
+
+You can add `run_eval` to start shell and set it True, if you want evaluation while training. And you can set argument option: `eval_dataset_path`, `save_best_ckpt`, `eval_start_epoch`, `eval_interval` when `run_eval` is True.
+
 ### Result
 
 - Training ResNet18 with CIFAR-10 dataset

model_zoo/official/cv/resnet/README_CN.md

@@ -143,7 +143,8 @@ bash run_eval_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH]
   ├── src
     ├── config.py                          # 参数配置
     ├── dataset.py                         # 数据预处理
-    ├── CrossEntropySmooth.py             # ImageNet2012数据集的损失定义
+    ├─  eval_callback.py                   # 训练时推理回调函数
+    ├── CrossEntropySmooth.py              # ImageNet2012数据集的损失定义
     ├── lr_generator.py                    # 生成每个步骤的学习率
     └── resnet.py                          # ResNet骨干网络，包括ResNet50、ResNet101和SE-ResNet50
   ├── eval.py                              # 评估网络
@@ -297,6 +298,10 @@ bash run_parameter_server_train.sh [resnet18|resnet50|resnet101] [cifar10|imagen
 bash run_parameter_server_train_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH]（可选）
 ```
 
+#### 训练时推理
+
+训练时推理需要在启动文件中添加`run_eval` 并设置为True。与此同时需要设置: `eval_dataset_path`, `save_best_ckpt`, `eval_start_epoch`, `eval_interval` 。
+
 ### 结果
 
 - 使用CIFAR-10数据集训练ResNet18

model_zoo/official/cv/resnet/src/eval_callback.py

@@ -0,0 +1,90 @@
+# Copyright 2021 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+"""Evaluation callback when training"""
+
+import os
+import stat
+from mindspore import save_checkpoint
+from mindspore import log as logger
+from mindspore.train.callback import Callback
+
+class EvalCallBack(Callback):
+    """
+    Evaluation callback when training.
+
+    Args:
+        eval_function (function): evaluation function.
+        eval_param_dict (dict): evaluation parameters' configure dict.
+        interval (int): run evaluation interval, default is 1.
+        eval_start_epoch (int): evaluation start epoch, default is 1.
+        save_best_ckpt (bool): Whether to save best checkpoint, default is True.
+        besk_ckpt_name (str): bast checkpoint name, default is `best.ckpt`.
+        metrics_name (str): evaluation metrics name, default is `acc`.
+
+    Returns:
+        None
+
+    Examples:
+        >>> EvalCallBack(eval_function, eval_param_dict)
+    """
+
+    def __init__(self, eval_function, eval_param_dict, interval=1, eval_start_epoch=1, save_best_ckpt=True,
+                 ckpt_directory="./", besk_ckpt_name="best.ckpt", metrics_name="acc"):
+        super(EvalCallBack, self).__init__()
+        self.eval_param_dict = eval_param_dict
+        self.eval_function = eval_function
+        self.eval_start_epoch = eval_start_epoch
+        if interval < 1:
+            raise ValueError("interval should >= 1.")
+        self.interval = interval
+        self.save_best_ckpt = save_best_ckpt
+        self.best_res = 0
+        self.best_epoch = 0
+        if not os.path.isdir(ckpt_directory):
+            os.makedirs(ckpt_directory)
+        self.bast_ckpt_path = os.path.join(ckpt_directory, besk_ckpt_name)
+        self.metrics_name = metrics_name
+
+    def remove_ckpoint_file(self, file_name):
+        """Remove the specified checkpoint file from this checkpoint manager and also from the directory."""
+        try:
+            os.chmod(file_name, stat.S_IWRITE)
+            os.remove(file_name)
+        except OSError:
+            logger.warning("OSError, failed to remove the older ckpt file %s.", file_name)
+        except ValueError:
+            logger.warning("ValueError, failed to remove the older ckpt file %s.", file_name)
+
+    def epoch_end(self, run_context):
+        """Callback when epoch end."""
+        cb_params = run_context.original_args()
+        cur_epoch = cb_params.cur_epoch_num
+        if cur_epoch >= self.eval_start_epoch and (cur_epoch - self.eval_start_epoch) % self.interval == 0:
+            res = self.eval_function(self.eval_param_dict)
+            print("epoch: {}, {}: {}".format(cur_epoch, self.metrics_name, res), flush=True)
+            if res >= self.best_res:
+                self.best_res = res
+                self.best_epoch = cur_epoch
+                print("update best result: {}".format(res), flush=True)
+                if self.save_best_ckpt:
+                    if os.path.exists(self.bast_ckpt_path):
+                        self.remove_ckpoint_file(self.bast_ckpt_path)
+                    save_checkpoint(cb_params.train_network, self.bast_ckpt_path)
+                    print("update best checkpoint at: {}".format(self.bast_ckpt_path), flush=True)
+
+    def end(self, run_context):
+        print("End training, the best {0} is: {1}, the best {0} epoch is {2}".format(self.metrics_name,
+                                                                                     self.best_res,
+                                                                                     self.best_epoch), flush=True)

model_zoo/official/cv/resnet/src/metric.py

@@ -0,0 +1,132 @@
+# Copyright 2021 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+"""evaluation metric."""
+
+from mindspore.communication.management import GlobalComm
+from mindspore.ops import operations as P
+import mindspore.nn as nn
+import mindspore.common.dtype as mstype
+
+
+class ClassifyCorrectCell(nn.Cell):
+    r"""
+    Cell that returns correct count of the prediction in classification network.
+    This Cell accepts a network as arguments.
+    It returns orrect count of the prediction to calculate the metrics.
+
+    Args:
+        network (Cell): The network Cell.
+
+    Inputs:
+        - **data** (Tensor) - Tensor of shape :math:`(N, \ldots)`.
+        - **label** (Tensor) - Tensor of shape :math:`(N, \ldots)`.
+
+    Outputs:
+        Tuple, containing a scalar correct count of the prediction
+
+    Examples:
+        >>> # For a defined network Net without loss function
+        >>> net = Net()
+        >>> eval_net = nn.ClassifyCorrectCell(net)
+    """
+
+    def __init__(self, network):
+        super(ClassifyCorrectCell, self).__init__(auto_prefix=False)
+        self._network = network
+        self.argmax = P.Argmax()
+        self.equal = P.Equal()
+        self.cast = P.Cast()
+        self.reduce_sum = P.ReduceSum()
+        self.allreduce = P.AllReduce(P.ReduceOp.SUM, GlobalComm.WORLD_COMM_GROUP)
+
+    def construct(self, data, label):
+        outputs = self._network(data)
+        y_pred = self.argmax(outputs)
+        y_pred = self.cast(y_pred, mstype.int32)
+        y_correct = self.equal(y_pred, label)
+        y_correct = self.cast(y_correct, mstype.float32)
+        y_correct = self.reduce_sum(y_correct)
+        total_correct = self.allreduce(y_correct)
+        return (total_correct,)
+
+
+class DistAccuracy(nn.Metric):
+    r"""
+    Calculates the accuracy for classification data in distributed mode.
+    The accuracy class creates two local variables, correct number and total number that are used to compute the
+    frequency with which predictions matches labels. This frequency is ultimately returned as the accuracy: an
+    idempotent operation that simply divides correct number by total number.
+
+    .. math::
+
+        \text{accuracy} =\frac{\text{true_positive} + \text{true_negative}}
+
+        {\text{true_positive} + \text{true_negative} + \text{false_positive} + \text{false_negative}}
+
+    Args:
+        eval_type (str): Metric to calculate the accuracy over a dataset, for classification (single-label).
+
+    Examples:
+        >>> y_correct = Tensor(np.array([20]))
+        >>> metric = nn.DistAccuracy(batch_size=3, device_num=8)
+        >>> metric.clear()
+        >>> metric.update(y_correct)
+        >>> accuracy = metric.eval()
+    """
+
+    def __init__(self, batch_size, device_num):
+        super(DistAccuracy, self).__init__()
+        self.clear()
+        self.batch_size = batch_size
+        self.device_num = device_num
+
+    def clear(self):
+        """Clears the internal evaluation result."""
+        self._correct_num = 0
+        self._total_num = 0
+
+    def update(self, *inputs):
+        """
+        Updates the internal evaluation result :math:`y_{pred}` and :math:`y`.
+
+        Args:
+            inputs: Input `y_correct`. `y_correct` is a `scalar Tensor`.
+                `y_correct` is the right prediction count that gathered from all devices
+                it's a scalar in float type
+
+        Raises:
+            ValueError: If the number of the input is not 1.
+        """
+
+        if len(inputs) != 1:
+            raise ValueError('Distribute accuracy needs 1 input (y_correct), but got {}'.format(len(inputs)))
+        y_correct = self._convert_data(inputs[0])
+        self._correct_num += y_correct
+        self._total_num += self.batch_size * self.device_num
+
+    def eval(self):
+        """
+        Computes the accuracy.
+
+        Returns:
+            Float, the computed result.
+
+        Raises:
+            RuntimeError: If the sample size is 0.
+        """
+
+        if self._total_num == 0:
+            raise RuntimeError('Accuracy can not be calculated, because the number of samples is 0.')
+        return self._correct_num / self._total_num

model_zoo/official/cv/resnet/train.py

@@ -31,9 +31,12 @@ from mindspore.common import set_seed
 from mindspore.parallel import set_algo_parameters
 import mindspore.nn as nn
 import mindspore.common.initializer as weight_init
+import mindspore.log as logger
 from src.lr_generator import get_lr, warmup_cosine_annealing_lr
 from src.CrossEntropySmooth import CrossEntropySmooth
 from src.config import cfg
+from src.eval_callback import EvalCallBack
+from src.metric import DistAccuracy, ClassifyCorrectCell
 
 parser = argparse.ArgumentParser(description='Image classification')
 parser.add_argument('--net', type=str, default=None, help='Resnet Model, resnet18, resnet50 or resnet101')
@@ -48,6 +51,15 @@ parser.add_argument('--pre_trained', type=str, default=None, help='Pretrained ch
 parser.add_argument('--parameter_server', type=ast.literal_eval, default=False, help='Run parameter server train')
 parser.add_argument("--filter_weight", type=ast.literal_eval, default=False,
                     help="Filter head weight parameters, default is False.")
+parser.add_argument("--run_eval", type=ast.literal_eval, default=False,
+                    help="Run evaluation when training, default is False.")
+parser.add_argument('--eval_dataset_path', type=str, default=None, help='Evaluation dataset path when run_eval is True')
+parser.add_argument("--save_best_ckpt", type=ast.literal_eval, default=True,
+                    help="Save best checkpoint when run_eval is True, default is True.")
+parser.add_argument("--eval_start_epoch", type=int, default=40,
+                    help="Evaluation start epoch when run_eval is True, default is 40.")
+parser.add_argument("--eval_interval", type=int, default=1,
+                    help="Evaluation interval when run_eval is True, default is 1.")
 args_opt = parser.parse_args()
 
 set_seed(1)
@@ -89,6 +101,12 @@ def filter_checkpoint_parameter_by_list(origin_dict, param_filter):
                 del origin_dict[key]
                 break
 
+def apply_eval(eval_param):
+    eval_model = eval_param["model"]
+    eval_ds = eval_param["dataset"]
+    metrics_name = eval_param["metrics_name"]
+    res = eval_model.eval(eval_ds)
+    return res[metrics_name]
 
 if __name__ == '__main__':
     target = args_opt.device_target
@@ -185,12 +203,16 @@ if __name__ == '__main__':
     else:
         loss = SoftmaxCrossEntropyWithLogits(sparse=True, reduction='mean')
     loss_scale = FixedLossScaleManager(config.loss_scale, drop_overflow_update=False)
-    model = Model(net, loss_fn=loss, optimizer=opt, loss_scale_manager=loss_scale, metrics={'acc'},
-                  amp_level="O2", keep_batchnorm_fp32=False)
+    dist_eval_network = ClassifyCorrectCell(net) if args_opt.run_distribute else None
+    metrics = {"acc"}
+    if args_opt.run_distribute:
+        metrics = {'acc': DistAccuracy(batch_size=config.batch_size, device_num=args_opt.device_num)}
+    model = Model(net, loss_fn=loss, optimizer=opt, loss_scale_manager=loss_scale, metrics=metrics,
+                  amp_level="O2", keep_batchnorm_fp32=False, eval_network=dist_eval_network)
     if (args_opt.net != "resnet101" and args_opt.net != "resnet50") or \
         args_opt.parameter_server or target == "CPU":
         ## fp32 training
-        model = Model(net, loss_fn=loss, optimizer=opt, metrics={'acc'})
+        model = Model(net, loss_fn=loss, optimizer=opt, metrics=metrics, eval_network=dist_eval_network)
     if cfg.optimizer == "Thor" and args_opt.dataset == "imagenet2012":
         from src.lr_generator import get_thor_damping
         damping = get_thor_damping(0, config.damping_init, config.damping_decay, 70, step_size)
@@ -201,6 +223,8 @@ if __name__ == '__main__':
                                                           loss_scale_manager=loss_scale, metrics={'acc'},
                                                           amp_level="O2", keep_batchnorm_fp32=False,
                                                           frequency=config.frequency)
+        args_opt.run_eval = False
+        logger.warning("Thor optimizer not support evaluation while training.")
 
     # define callbacks
     time_cb = TimeMonitor(data_size=step_size)
@@ -211,7 +235,17 @@ if __name__ == '__main__':
                                      keep_checkpoint_max=config.keep_checkpoint_max)
         ckpt_cb = ModelCheckpoint(prefix="resnet", directory=ckpt_save_dir, config=config_ck)
         cb += [ckpt_cb]
-
+    if args_opt.run_eval:
+        if args_opt.eval_dataset_path is None or (not os.path.isdir(args_opt.eval_dataset_path)):
+            raise ValueError("{} is not a existing path.".format(args_opt.eval_dataset_path))
+        eval_dataset = create_dataset(dataset_path=args_opt.eval_dataset_path, do_train=False,
+                                      batch_size=config.batch_size, target=target)
+        eval_param_dict = {"model": model, "dataset": eval_dataset, "metrics_name": "acc"}
+        eval_cb = EvalCallBack(apply_eval, eval_param_dict, interval=args_opt.eval_interval,
+                               eval_start_epoch=args_opt.eval_start_epoch, save_best_ckpt=True,
+                               ckpt_directory=ckpt_save_dir, besk_ckpt_name="best_acc.ckpt",
+                               metrics_name="acc")
+        cb += [eval_cb]
     # train model
     if args_opt.net == "se-resnet50":
         config.epoch_size = config.train_epoch_size

model_zoo/official/cv/ssd/README.md

@@ -123,8 +123,8 @@ Dataset used: [COCO2017](<http://images.cocodataset.org/>)
 
 ### Prepare the model
 
-1. Chose the model by changing the `using_model` in `src/confgi.py`. The optional models are: `ssd300`, `ssd_mobilenet_v1_fpn`.
-2. Change the dataset config in the corresponding config. `src/config_ssd300.py` or `src/config_ssd_mobilenet_v1_fpn.py`.
+1. Chose the model by changing the `using_model` in `src/confgi.py`. The optional models are: `ssd300`, `ssd_mobilenet_v1_fpn`, `ssd_mobilenet_v1_fpn`, `ssd_resnet50_fpn`.
+2. Change the dataset config in the corresponding config. `src/config_ssd300.py`, `src/config_ssd_mobilenet_v1_fpn.py`, `src/config_ssd_resnet50_fpn.py`, `src/config_ssd_vgg16.py`.
 3. If you are running with `ssd_mobilenet_v1_fpn`, you need a pretrained model for `mobilenet_v1`. Set the checkpoint path to `feature_extractor_base_param` in `src/config_ssd_mobilenet_v1_fpn.py`. For more detail about training mobilnet_v1, please refer to the mobilenetv1 model.
 
 ### Run the scripts
@@ -201,6 +201,7 @@ Then you can run everything just like on ascend.
     ├─ src
       ├─ __init__.py                  # init file
       ├─ box_utils.py                 # bbox utils
+      ├─ eval_callback.py             # evaluation callback when training
       ├─ eval_utils.py                # metrics utils
       ├─ config.py                    # total config
       ├─ dataset.py                   # create dataset and process dataset
@@ -229,6 +230,10 @@ Then you can run everything just like on ascend.
     "loss_scale": 1024                               # Loss scale
     "filter_weight": False                           # Load parameters in head layer or not. If the class numbers of train dataset is different from the class numbers in pre_trained checkpoint, please set True.
     "freeze_layer": "none"                           # Freeze the backbone parameters or not, support none and backbone.
+    "run_eval": False                                # Run evaluation when training
+    "save_best_ckpt": True                           # Save best checkpoint when run_eval is True
+    "eval_start_epoch": 40                           # Evaluation start epoch when run_eval is True
+    "eval_interval": 1                               # valuation interval when run_eval is True
 
     "class_num": 81                                  # Dataset class number
     "image_shape": [300, 300]                        # Image height and width used as input to the model
@@ -311,6 +316,10 @@ epoch time: 150753.701, per step time: 329.157
 ...
 ```
 
+#### Evaluation while training
+
+You can add `run_eval` to start shell and set it True, if you want evaluation while training. And you can set argument option: `save_best_ckpt`, `eval_start_epoch`, `eval_interval` when `run_eval` is True.
+
 #### Transfer Training
 
 You can train your own model based on either pretrained classification model or pretrained detection model. You can perform transfer training by following steps.

model_zoo/official/cv/ssd/eval.py

@@ -17,14 +17,12 @@
 
 import os
 import argparse
-import time
-import numpy as np
 from mindspore import context, Tensor
 from mindspore.train.serialization import load_checkpoint, load_param_into_net
 from src.ssd import SSD300, SsdInferWithDecoder, ssd_mobilenet_v2, ssd_mobilenet_v1_fpn, ssd_resnet50_fpn, ssd_vgg16
 from src.dataset import create_ssd_dataset, create_mindrecord
 from src.config import config
-from src.eval_utils import metrics
+from src.eval_utils import apply_eval
 from src.box_utils import default_boxes
 
 def ssd_eval(dataset_path, ckpt_path, anno_json):
@@ -50,31 +48,12 @@ def ssd_eval(dataset_path, ckpt_path, anno_json):
     load_param_into_net(net, param_dict)
 
     net.set_train(False)
-    i = batch_size
     total = ds.get_dataset_size() * batch_size
-    start = time.time()
-    pred_data = []
     print("\n========================================\n")
     print("total images num: ", total)
     print("Processing, please wait a moment.")
-    for data in ds.create_dict_iterator(output_numpy=True, num_epochs=1):
-        img_id = data['img_id']
-        img_np = data['image']
-        image_shape = data['image_shape']
-
-        output = net(Tensor(img_np))
-        for batch_idx in range(img_np.shape[0]):
-            pred_data.append({"boxes": output[0].asnumpy()[batch_idx],
-                              "box_scores": output[1].asnumpy()[batch_idx],
-                              "img_id": int(np.squeeze(img_id[batch_idx])),
-                              "image_shape": image_shape[batch_idx]})
-        percent = round(i / total * 100., 2)
-
-        print(f'    {str(percent)} [{i}/{total}]', end='\r')
-        i += batch_size
-    cost_time = int((time.time() - start) * 1000)
-    print(f'    100% [{total}/{total}] cost {cost_time} ms')
-    mAP = metrics(pred_data, anno_json)
+    eval_param_dict = {"net": net, "dataset": ds, "anno_json": anno_json}
+    mAP = apply_eval(eval_param_dict)
     print("\n========================================\n")
     print(f"mAP: {mAP}")
 

model_zoo/official/cv/ssd/src/eval_callback.py

@@ -0,0 +1,90 @@
+# Copyright 2021 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+"""Evaluation callback when training"""
+
+import os
+import stat
+from mindspore import save_checkpoint
+from mindspore import log as logger
+from mindspore.train.callback import Callback
+
+class EvalCallBack(Callback):
+    """
+    Evaluation callback when training.
+
+    Args:
+        eval_function (function): evaluation function.
+        eval_param_dict (dict): evaluation parameters' configure dict.
+        interval (int): run evaluation interval, default is 1.
+        eval_start_epoch (int): evaluation start epoch, default is 1.
+        save_best_ckpt (bool): Whether to save best checkpoint, default is True.
+        besk_ckpt_name (str): bast checkpoint name, default is `best.ckpt`.
+        metrics_name (str): evaluation metrics name, default is `acc`.
+
+    Returns:
+        None
+
+    Examples:
+        >>> EvalCallBack(eval_function, eval_param_dict)
+    """
+
+    def __init__(self, eval_function, eval_param_dict, interval=1, eval_start_epoch=1, save_best_ckpt=True,
+                 ckpt_directory="./", besk_ckpt_name="best.ckpt", metrics_name="acc"):
+        super(EvalCallBack, self).__init__()
+        self.eval_param_dict = eval_param_dict
+        self.eval_function = eval_function
+        self.eval_start_epoch = eval_start_epoch
+        if interval < 1:
+            raise ValueError("interval should >= 1.")
+        self.interval = interval
+        self.save_best_ckpt = save_best_ckpt
+        self.best_res = 0
+        self.best_epoch = 0
+        if not os.path.isdir(ckpt_directory):
+            os.makedirs(ckpt_directory)
+        self.bast_ckpt_path = os.path.join(ckpt_directory, besk_ckpt_name)
+        self.metrics_name = metrics_name
+
+    def remove_ckpoint_file(self, file_name):
+        """Remove the specified checkpoint file from this checkpoint manager and also from the directory."""
+        try:
+            os.chmod(file_name, stat.S_IWRITE)
+            os.remove(file_name)
+        except OSError:
+            logger.warning("OSError, failed to remove the older ckpt file %s.", file_name)
+        except ValueError:
+            logger.warning("ValueError, failed to remove the older ckpt file %s.", file_name)
+
+    def epoch_end(self, run_context):
+        """Callback when epoch end."""
+        cb_params = run_context.original_args()
+        cur_epoch = cb_params.cur_epoch_num
+        if cur_epoch >= self.eval_start_epoch and (cur_epoch - self.eval_start_epoch) % self.interval == 0:
+            res = self.eval_function(self.eval_param_dict)
+            print("epoch: {}, {}: {}".format(cur_epoch, self.metrics_name, res), flush=True)
+            if res >= self.best_res:
+                self.best_res = res
+                self.best_epoch = cur_epoch
+                print("update best result: {}".format(res), flush=True)
+                if self.save_best_ckpt:
+                    if os.path.exists(self.bast_ckpt_path):
+                        self.remove_ckpoint_file(self.bast_ckpt_path)
+                    save_checkpoint(cb_params.train_network, self.bast_ckpt_path)
+                    print("update best checkpoint at: {}".format(self.bast_ckpt_path), flush=True)
+
+    def end(self, run_context):
+        print("End training, the best {0} is: {1}, the best {0} epoch is {2}".format(self.metrics_name,
+                                                                                     self.best_res,
+                                                                                     self.best_epoch), flush=True)

model_zoo/official/cv/ssd/src/eval_utils.py

@@ -16,8 +16,28 @@
 
 import json
 import numpy as np
+from mindspore import Tensor
 from .config import config
 
+def apply_eval(eval_param_dict):
+    net = eval_param_dict["net"]
+    net.set_train(False)
+    ds = eval_param_dict["dataset"]
+    anno_json = eval_param_dict["anno_json"]
+    pred_data = []
+    for data in ds.create_dict_iterator(output_numpy=True, num_epochs=1):
+        img_id = data['img_id']
+        img_np = data['image']
+        image_shape = data['image_shape']
+
+        output = net(Tensor(img_np))
+        for batch_idx in range(img_np.shape[0]):
+            pred_data.append({"boxes": output[0].asnumpy()[batch_idx],
+                              "box_scores": output[1].asnumpy()[batch_idx],
+                              "img_id": int(np.squeeze(img_id[batch_idx])),
+                              "image_shape": image_shape[batch_idx]})
+    mAP = metrics(pred_data, anno_json)
+    return mAP
 
 def apply_nms(all_boxes, all_scores, thres, max_boxes):
     """Apply NMS to bboxes."""

model_zoo/official/cv/ssd/train.py

@@ -15,6 +15,7 @@
 
 """Train SSD and get checkpoint files."""
 
+import os
 import argparse
 import ast
 import mindspore.nn as nn
@@ -25,11 +26,15 @@ from mindspore.train import Model
 from mindspore.context import ParallelMode
 from mindspore.train.serialization import load_checkpoint, load_param_into_net
 from mindspore.common import set_seed, dtype
-from src.ssd import SSD300, SSDWithLossCell, TrainingWrapper, ssd_mobilenet_v2, ssd_mobilenet_v1_fpn, ssd_resnet50_fpn, ssd_vgg16
+from src.ssd import SSD300, SsdInferWithDecoder, SSDWithLossCell, TrainingWrapper, ssd_mobilenet_v2,\
+    ssd_mobilenet_v1_fpn, ssd_resnet50_fpn, ssd_vgg16
 from src.config import config
 from src.dataset import create_ssd_dataset, create_mindrecord
 from src.lr_schedule import get_lr
 from src.init_params import init_net_param, filter_checkpoint_parameter_by_list
+from src.eval_callback import EvalCallBack
+from src.eval_utils import apply_eval
+from src.box_utils import default_boxes
 
 set_seed(1)
 
@@ -57,6 +62,14 @@ def get_args():
     parser.add_argument('--freeze_layer', type=str, default="none", choices=["none", "backbone"],
                         help="freeze the weights of network, support freeze the backbone's weights, "
                              "default is not freezing.")
+    parser.add_argument("--run_eval", type=ast.literal_eval, default=False,
+                        help="Run evaluation when training, default is False.")
+    parser.add_argument("--save_best_ckpt", type=ast.literal_eval, default=True,
+                        help="Save best checkpoint when run_eval is True, default is True.")
+    parser.add_argument("--eval_start_epoch", type=int, default=40,
+                        help="Evaluation start epoch when run_eval is True, default is 40.")
+    parser.add_argument("--eval_interval", type=int, default=1,
+                        help="Evaluation interval when run_eval is True, default is 1.")
     args_opt = parser.parse_args()
     return args_opt
 
@@ -170,8 +183,25 @@ def main():
                           config.momentum, config.weight_decay, loss_scale)
         net = TrainingWrapper(net, opt, loss_scale)
 
-
     callback = [TimeMonitor(data_size=dataset_size), LossMonitor(), ckpoint_cb]
+    if args_opt.run_eval:
+        eval_net = SsdInferWithDecoder(ssd, Tensor(default_boxes), config)
+        eval_net.set_train(False)
+        mindrecord_file = create_mindrecord(args_opt.dataset, "ssd_eval.mindrecord", False)
+        eval_dataset = create_ssd_dataset(mindrecord_file, batch_size=args_opt.batch_size, repeat_num=1,
+                                          is_training=False, use_multiprocessing=False)
+        if args_opt.dataset == "coco":
+            anno_json = os.path.join(config.coco_root, config.instances_set.format(config.val_data_type))
+        elif args_opt.dataset == "voc":
+            anno_json = os.path.join(config.voc_root, config.voc_json)
+        else:
+            raise ValueError('SSD eval only support dataset mode is coco and voc!')
+        eval_param_dict = {"net": eval_net, "dataset": eval_dataset, "anno_json": anno_json}
+        eval_cb = EvalCallBack(apply_eval, eval_param_dict, interval=args_opt.eval_interval,
+                               eval_start_epoch=args_opt.eval_start_epoch, save_best_ckpt=True,
+                               ckpt_directory=save_ckpt_path, besk_ckpt_name="best_map.ckpt",
+                               metrics_name="mAP")
+        callback.append(eval_cb)
     model = Model(net)
     dataset_sink_mode = False
     if args_opt.mode == "sink" and args_opt.run_platform != "CPU":

model_zoo/official/cv/unet/README.md

@@ -128,6 +128,7 @@ Then you can run everything just like on ascend.
         │   ├──config.py                    // parameter configuration
         │   ├──data_loader.py               // creating dataset
         │   ├──loss.py                      // loss
+        │   ├──eval_callback.py             // evaluation callback while training
         │   ├──utils.py                     // General components (callback function)
         │   ├──unet_medical                 // Unet medical architecture
                 ├──__init__.py              // init file
@@ -168,6 +169,11 @@ Parameters for both training and evaluation can be set in config.py
   'resume_ckpt': './',                # pretrain model path
   'transfer_training': False          # whether do transfer training
   'filter_weight': ["final.weight"]   # weight name to filter while doing transfer training
+  'run_eval': False                   # Run evaluation when training
+  'save_best_ckpt': True              # Save best checkpoint when run_eval is True
+  'eval_start_epoch': 0               # Evaluation start epoch when run_eval is True
+  'eval_interval': 1                  # valuation interval when run_eval is True
+
   ```
 
 - config for Unet++, cell nuclei dataset
@@ -193,6 +199,10 @@ Parameters for both training and evaluation can be set in config.py
   'resume_ckpt': './',                # pretrain model path
   'transfer_training': False          # whether do transfer training
   'filter_weight': ['final1.weight', 'final2.weight', 'final3.weight', 'final4.weight']  # weight name to filter while doing transfer training
+  'run_eval': False                   # Run evaluation when training
+  'save_best_ckpt': True              # Save best checkpoint when run_eval is True
+  'eval_start_epoch': 0               # Evaluation start epoch when run_eval is True
+  'eval_interval': 1                  # valuation interval when run_eval is True
   ```
 
 ## [Training Process](#contents)
@@ -245,6 +255,10 @@ step: 299, loss is 0.20551169, fps is 58.4039329983891
 step: 300, loss is 0.18949677, fps is 57.63118508760329
 ```
 
+#### Evaluation while training
+
+You can add `run_eval` to start shell and set it True, if you want evaluation while training. And you can set argument option: `save_best_ckpt`, `eval_start_epoch`, `eval_interval`, `eval_metrics` when `run_eval` is True.
+
 ## [Evaluation Process](#contents)
 
 ### Evaluation

model_zoo/official/cv/unet/README_CN.md

@@ -132,6 +132,7 @@ bash scripts/docker_start.sh unet:20.1.0 [DATA_DIR] [MODEL_DIR]
         │   ├──config.py                    // 参数配置
         │   ├──data_loader.py               // 数据处理
         │   ├──loss.py                      // 损失函数
+        │   ├─  eval_callback.py            // 训练时推理回调函数
         │   ├──utils.py                     // 通用组件（回调函数）
         │   ├──unet_medical                 // 医学图像处理Unet结构
                 ├──__init__.py
@@ -247,6 +248,10 @@ step: 299, loss is 0.20551169, fps is 58.4039329983891
 step: 300, loss is 0.18949677, fps is 57.63118508760329
 ```
 
+#### 训练时推理
+
+训练时推理需要在启动文件中添加`run_eval` 并设置为True。与此同时需要设置: `save_best_ckpt`, `eval_start_epoch`, `eval_interval`, `eval_metrics` 。
+
 ## 评估过程
 
 ### 评估

model_zoo/official/cv/unet/eval.py

@@ -16,10 +16,6 @@
 import os
 import argparse
 import logging
-import cv2
-import numpy as np
-import mindspore.nn as nn
-import mindspore.ops.operations as F
 from mindspore import context, Model
 from mindspore.train.serialization import load_checkpoint, load_param_into_net
 
@@ -27,76 +23,11 @@ from src.data_loader import create_dataset, create_cell_nuclei_dataset
 from src.unet_medical import UNetMedical
 from src.unet_nested import NestedUNet, UNet
 from src.config import cfg_unet
-from src.utils import UnetEval
+from src.utils import UnetEval, TempLoss, dice_coeff
 
 device_id = int(os.getenv('DEVICE_ID'))
 context.set_context(mode=context.GRAPH_MODE, device_target="Ascend", save_graphs=False, device_id=device_id)
 
-
-class TempLoss(nn.Cell):
-    """A temp loss cell."""
-    def __init__(self):
-        super(TempLoss, self).__init__()
-        self.identity = F.identity()
-    def construct(self, logits, label):
-        return self.identity(logits)
-
-
-class dice_coeff(nn.Metric):
-    def __init__(self):
-        super(dice_coeff, self).__init__()
-        self.clear()
-    def clear(self):
-        self._dice_coeff_sum = 0
-        self._iou_sum = 0
-        self._samples_num = 0
-
-    def update(self, *inputs):
-        if len(inputs) != 2:
-            raise ValueError('Need 2 inputs ((y_softmax, y_argmax), y), but got {}'.format(len(inputs)))
-        y = self._convert_data(inputs[1])
-        self._samples_num += y.shape[0]
-        y = y.transpose(0, 2, 3, 1)
-        b, h, w, c = y.shape
-        if b != 1:
-            raise ValueError('Batch size should be 1 when in evaluation.')
-        y = y.reshape((h, w, c))
-        if cfg_unet["eval_activate"].lower() == "softmax":
-            y_softmax = np.squeeze(self._convert_data(inputs[0][0]), axis=0)
-            if cfg_unet["eval_resize"]:
-                y_pred = []
-                for i in range(cfg_unet["num_classes"]):
-                    y_pred.append(cv2.resize(np.uint8(y_softmax[:, :, i] * 255), (w, h)) / 255)
-                y_pred = np.stack(y_pred, axis=-1)
-            else:
-                y_pred = y_softmax
-        elif cfg_unet["eval_activate"].lower() == "argmax":
-            y_argmax = np.squeeze(self._convert_data(inputs[0][1]), axis=0)
-            y_pred = []
-            for i in range(cfg_unet["num_classes"]):
-                if cfg_unet["eval_resize"]:
-                    y_pred.append(cv2.resize(np.uint8(y_argmax == i), (w, h), interpolation=cv2.INTER_NEAREST))
-                else:
-                    y_pred.append(np.float32(y_argmax == i))
-            y_pred = np.stack(y_pred, axis=-1)
-        else:
-            raise ValueError('config eval_activate should be softmax or argmax.')
-        y_pred = y_pred.astype(np.float32)
-        inter = np.dot(y_pred.flatten(), y.flatten())
-        union = np.dot(y_pred.flatten(), y_pred.flatten()) + np.dot(y.flatten(), y.flatten())
-
-        single_dice_coeff = 2*float(inter)/float(union+1e-6)
-        single_iou = single_dice_coeff / (2 - single_dice_coeff)
-        print("single dice coeff is: {}, IOU is: {}".format(single_dice_coeff, single_iou))
-        self._dice_coeff_sum += single_dice_coeff
-        self._iou_sum += single_iou
-
-    def eval(self):
-        if self._samples_num == 0:
-            raise RuntimeError('Total samples num must not be 0.')
-        return (self._dice_coeff_sum / float(self._samples_num), self._iou_sum / float(self._samples_num))
-
-
 def test_net(data_dir,
              ckpt_path,
              cross_valid_ind=1,
@@ -119,7 +50,7 @@ def test_net(data_dir,
     else:
         _, valid_dataset = create_dataset(data_dir, 1, 1, False, cross_valid_ind, False,
                                           do_crop=cfg['crop'], img_size=cfg['img_size'])
-    model = Model(net, loss_fn=TempLoss(), metrics={"dice_coeff": dice_coeff()})
+    model = Model(net, loss_fn=TempLoss(), metrics={"dice_coeff": dice_coeff(cfg_unet)})
 
     print("============== Starting Evaluating ============")
     eval_score = model.eval(valid_dataset, dataset_sink_mode=False)["dice_coeff"]

model_zoo/official/cv/unet/src/eval_callback.py

@@ -0,0 +1,90 @@
+# Copyright 2021 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+"""Evaluation callback when training"""
+
+import os
+import stat
+from mindspore import save_checkpoint
+from mindspore import log as logger
+from mindspore.train.callback import Callback
+
+class EvalCallBack(Callback):
+    """
+    Evaluation callback when training.
+
+    Args:
+        eval_function (function): evaluation function.
+        eval_param_dict (dict): evaluation parameters' configure dict.
+        interval (int): run evaluation interval, default is 1.
+        eval_start_epoch (int): evaluation start epoch, default is 1.
+        save_best_ckpt (bool): Whether to save best checkpoint, default is True.
+        besk_ckpt_name (str): bast checkpoint name, default is `best.ckpt`.
+        metrics_name (str): evaluation metrics name, default is `acc`.
+
+    Returns:
+        None
+
+    Examples:
+        >>> EvalCallBack(eval_function, eval_param_dict)
+    """
+
+    def __init__(self, eval_function, eval_param_dict, interval=1, eval_start_epoch=1, save_best_ckpt=True,
+                 ckpt_directory="./", besk_ckpt_name="best.ckpt", metrics_name="acc"):
+        super(EvalCallBack, self).__init__()
+        self.eval_param_dict = eval_param_dict
+        self.eval_function = eval_function
+        self.eval_start_epoch = eval_start_epoch
+        if interval < 1:
+            raise ValueError("interval should >= 1.")
+        self.interval = interval
+        self.save_best_ckpt = save_best_ckpt
+        self.best_res = 0
+        self.best_epoch = 0
+        if not os.path.isdir(ckpt_directory):
+            os.makedirs(ckpt_directory)
+        self.bast_ckpt_path = os.path.join(ckpt_directory, besk_ckpt_name)
+        self.metrics_name = metrics_name
+
+    def remove_ckpoint_file(self, file_name):
+        """Remove the specified checkpoint file from this checkpoint manager and also from the directory."""
+        try:
+            os.chmod(file_name, stat.S_IWRITE)
+            os.remove(file_name)
+        except OSError:
+            logger.warning("OSError, failed to remove the older ckpt file %s.", file_name)
+        except ValueError:
+            logger.warning("ValueError, failed to remove the older ckpt file %s.", file_name)
+
+    def epoch_end(self, run_context):
+        """Callback when epoch end."""
+        cb_params = run_context.original_args()
+        cur_epoch = cb_params.cur_epoch_num
+        if cur_epoch >= self.eval_start_epoch and (cur_epoch - self.eval_start_epoch) % self.interval == 0:
+            res = self.eval_function(self.eval_param_dict)
+            print("epoch: {}, {}: {}".format(cur_epoch, self.metrics_name, res), flush=True)
+            if res >= self.best_res:
+                self.best_res = res
+                self.best_epoch = cur_epoch
+                print("update best result: {}".format(res), flush=True)
+                if self.save_best_ckpt:
+                    if os.path.exists(self.bast_ckpt_path):
+                        self.remove_ckpoint_file(self.bast_ckpt_path)
+                    save_checkpoint(cb_params.train_network, self.bast_ckpt_path)
+                    print("update best checkpoint at: {}".format(self.bast_ckpt_path), flush=True)
+
+    def end(self, run_context):
+        print("End training, the best {0} is: {1}, the best {0} epoch is {2}".format(self.metrics_name,
+                                                                                     self.best_res,
+                                                                                     self.best_epoch), flush=True)

model_zoo/official/cv/unet/src/loss.py

@@ -41,7 +41,7 @@ class MultiCrossEntropyWithLogits(nn.Cell):
     def __init__(self):
         super(MultiCrossEntropyWithLogits, self).__init__()
         self.loss = CrossEntropyWithLogits()
-        self.squeeze = F.Squeeze()
+        self.squeeze = F.Squeeze(axis=0)
 
     def construct(self, logits, label):
         total_loss = 0

model_zoo/official/cv/unet/src/utils.py

@@ -14,6 +14,7 @@
 # ============================================================================
 
 import time
+import cv2
 import numpy as np
 from PIL import Image
 from mindspore import nn
@@ -25,20 +26,100 @@ class UnetEval(nn.Cell):
     """
     Add Unet evaluation activation.
     """
-    def __init__(self, net):
+    def __init__(self, net, need_slice=False):
         super(UnetEval, self).__init__()
         self.net = net
+        self.need_slice = need_slice
         self.transpose = ops.Transpose()
         self.softmax = ops.Softmax(axis=-1)
         self.argmax = ops.Argmax(axis=-1)
+        self.squeeze = ops.Squeeze(axis=0)
 
     def construct(self, x):
         out = self.net(x)
+        if self.need_slice:
+            out = self.squeeze(out[-1:])
         out = self.transpose(out, (0, 2, 3, 1))
         softmax_out = self.softmax(out)
         argmax_out = self.argmax(out)
         return (softmax_out, argmax_out)
 
+class TempLoss(nn.Cell):
+    """A temp loss cell."""
+    def __init__(self):
+        super(TempLoss, self).__init__()
+        self.identity = ops.identity()
+    def construct(self, logits, label):
+        return self.identity(logits)
+
+def apply_eval(eval_param_dict):
+    """run Evaluation"""
+    model = eval_param_dict["model"]
+    dataset = eval_param_dict["dataset"]
+    metrics_name = eval_param_dict["metrics_name"]
+    index = 0 if metrics_name == "dice_coeff" else 1
+    eval_score = model.eval(dataset, dataset_sink_mode=False)[metrics_name][index]
+    return eval_score
+
+class dice_coeff(nn.Metric):
+    """Unet Metric, return dice coefficient and IOU."""
+    def __init__(self, cfg_unet, print_res=True):
+        super(dice_coeff, self).__init__()
+        self.clear()
+        self.cfg_unet = cfg_unet
+        self.print_res = print_res
+
+    def clear(self):
+        self._dice_coeff_sum = 0
+        self._iou_sum = 0
+        self._samples_num = 0
+
+    def update(self, *inputs):
+        if len(inputs) != 2:
+            raise ValueError('Need 2 inputs ((y_softmax, y_argmax), y), but got {}'.format(len(inputs)))
+        y = self._convert_data(inputs[1])
+        self._samples_num += y.shape[0]
+        y = y.transpose(0, 2, 3, 1)
+        b, h, w, c = y.shape
+        if b != 1:
+            raise ValueError('Batch size should be 1 when in evaluation.')
+        y = y.reshape((h, w, c))
+        if self.cfg_unet["eval_activate"].lower() == "softmax":
+            y_softmax = np.squeeze(self._convert_data(inputs[0][0]), axis=0)
+            if self.cfg_unet["eval_resize"]:
+                y_pred = []
+                for i in range(self.cfg_unet["num_classes"]):
+                    y_pred.append(cv2.resize(np.uint8(y_softmax[:, :, i] * 255), (w, h)) / 255)
+                y_pred = np.stack(y_pred, axis=-1)
+            else:
+                y_pred = y_softmax
+        elif self.cfg_unet["eval_activate"].lower() == "argmax":
+            y_argmax = np.squeeze(self._convert_data(inputs[0][1]), axis=0)
+            y_pred = []
+            for i in range(self.cfg_unet["num_classes"]):
+                if self.cfg_unet["eval_resize"]:
+                    y_pred.append(cv2.resize(np.uint8(y_argmax == i), (w, h), interpolation=cv2.INTER_NEAREST))
+                else:
+                    y_pred.append(np.float32(y_argmax == i))
+            y_pred = np.stack(y_pred, axis=-1)
+        else:
+            raise ValueError('config eval_activate should be softmax or argmax.')
+        y_pred = y_pred.astype(np.float32)
+        inter = np.dot(y_pred.flatten(), y.flatten())
+        union = np.dot(y_pred.flatten(), y_pred.flatten()) + np.dot(y.flatten(), y.flatten())
+
+        single_dice_coeff = 2 * float(inter) / float(union+1e-6)
+        single_iou = single_dice_coeff / (2 - single_dice_coeff)
+        if self.print_res:
+            print("single dice coeff is: {}, IOU is: {}".format(single_dice_coeff, single_iou))
+        self._dice_coeff_sum += single_dice_coeff
+        self._iou_sum += single_iou
+
+    def eval(self):
+        if self._samples_num == 0:
+            raise RuntimeError('Total samples num must not be 0.')
+        return (self._dice_coeff_sum / float(self._samples_num), self._iou_sum / float(self._samples_num))
+
 class StepLossTimeMonitor(Callback):
 
     def __init__(self, batch_size, per_print_times=1):

model_zoo/official/cv/unet/train.py

@@ -30,23 +30,25 @@ from src.unet_medical import UNetMedical
 from src.unet_nested import NestedUNet, UNet
 from src.data_loader import create_dataset, create_cell_nuclei_dataset
 from src.loss import CrossEntropyWithLogits, MultiCrossEntropyWithLogits
-from src.utils import StepLossTimeMonitor, filter_checkpoint_parameter_by_list
+from src.utils import StepLossTimeMonitor, UnetEval, TempLoss, apply_eval, filter_checkpoint_parameter_by_list, dice_coeff
 from src.config import cfg_unet
+from src.eval_callback import EvalCallBack
 
 device_id = int(os.getenv('DEVICE_ID'))
 context.set_context(mode=context.GRAPH_MODE, device_target="Ascend", save_graphs=False, device_id=device_id)
 
 mindspore.set_seed(1)
 
-def train_net(data_dir,
+def train_net(args_opt,
               cross_valid_ind=1,
               epochs=400,
               batch_size=16,
               lr=0.0001,
-              run_distribute=False,
               cfg=None):
     rank = 0
     group_size = 1
+    data_dir = args_opt.data_url
+    run_distribute = args_opt.run_distribute
     if run_distribute:
         init()
         group_size = get_group_size()
@@ -55,12 +57,13 @@ def train_net(data_dir,
         context.set_auto_parallel_context(parallel_mode=parallel_mode,
                                           device_num=group_size,
                                           gradients_mean=False)
-
+    need_slice = False
     if cfg['model'] == 'unet_medical':
         net = UNetMedical(n_channels=cfg['num_channels'], n_classes=cfg['num_classes'])
     elif cfg['model'] == 'unet_nested':
         net = NestedUNet(in_channel=cfg['num_channels'], n_class=cfg['num_classes'], use_deconv=cfg['use_deconv'],
                          use_bn=cfg['use_bn'], use_ds=cfg['use_ds'])
+        need_slice = cfg['use_ds']
     elif cfg['model'] == 'unet_simple':
         net = UNet(in_channel=cfg['num_channels'], n_class=cfg['num_classes'])
     else:
@@ -83,12 +86,15 @@ def train_net(data_dir,
         train_dataset = create_cell_nuclei_dataset(data_dir, cfg['img_size'], repeat, batch_size,
                                                    is_train=True, augment=True, split=0.8, rank=rank,
                                                    group_size=group_size)
+        valid_dataset = create_cell_nuclei_dataset(data_dir, cfg['img_size'], 1, 1, is_train=False,
+                                                   eval_resize=cfg["eval_resize"], split=0.8,
+                                                   python_multiprocessing=False)
     else:
         repeat = epochs
         dataset_sink_mode = False
         per_print_times = 1
-        train_dataset, _ = create_dataset(data_dir, repeat, batch_size, True, cross_valid_ind, run_distribute,
-                                          cfg["crop"], cfg['img_size'])
+        train_dataset, valid_dataset = create_dataset(data_dir, repeat, batch_size, True, cross_valid_ind,
+                                                      run_distribute, cfg["crop"], cfg['img_size'])
     train_data_size = train_dataset.get_dataset_size()
     print("dataset length is:", train_data_size)
     ckpt_config = CheckpointConfig(save_checkpoint_steps=train_data_size,
@@ -106,6 +112,15 @@ def train_net(data_dir,
 
     print("============== Starting Training ==============")
     callbacks = [StepLossTimeMonitor(batch_size=batch_size, per_print_times=per_print_times), ckpoint_cb]
+    if args_opt.run_eval:
+        eval_model = Model(UnetEval(net, need_slice=need_slice), loss_fn=TempLoss(),
+                           metrics={"dice_coeff": dice_coeff(cfg_unet, False)})
+        eval_param_dict = {"model": eval_model, "dataset": valid_dataset, "metrics_name": args_opt.eval_metrics}
+        eval_cb = EvalCallBack(apply_eval, eval_param_dict, interval=args_opt.eval_interval,
+                               eval_start_epoch=args_opt.eval_start_epoch, save_best_ckpt=True,
+                               ckpt_directory='./ckpt_{}/'.format(device_id), besk_ckpt_name="best.ckpt",
+                               metrics_name=args_opt.eval_metrics)
+        callbacks.append(eval_cb)
     model.train(int(epochs / repeat), train_dataset, callbacks=callbacks, dataset_sink_mode=dataset_sink_mode)
     print("============== End Training ==============")
 
@@ -117,6 +132,17 @@ def get_args():
                         help='data directory')
     parser.add_argument('-t', '--run_distribute', type=ast.literal_eval,
                         default=False, help='Run distribute, default: false.')
+    parser.add_argument("--run_eval", type=ast.literal_eval, default=False,
+                        help="Run evaluation when training, default is False.")
+    parser.add_argument("--save_best_ckpt", type=ast.literal_eval, default=True,
+                        help="Save best checkpoint when run_eval is True, default is True.")
+    parser.add_argument("--eval_start_epoch", type=int, default=0,
+                        help="Evaluation start epoch when run_eval is True, default is 0.")
+    parser.add_argument("--eval_interval", type=int, default=1,
+                        help="Evaluation interval when run_eval is True, default is 1.")
+    parser.add_argument("--eval_metrics", type=str, default="dice_coeff", choices=("dice_coeff", "iou"),
+                        help="Evaluation metrics when run_eval is True, support [dice_coeff, iou], "
+                             "default is dice_coeff.")
 
     return parser.parse_args()
 
@@ -127,10 +153,9 @@ if __name__ == '__main__':
     print("Training setting:", args)
 
     epoch_size = cfg_unet['epochs'] if not args.run_distribute else cfg_unet['distribute_epochs']
-    train_net(data_dir=args.data_url,
+    train_net(args_opt=args,
               cross_valid_ind=cfg_unet['cross_valid_ind'],
               epochs=epoch_size,
               batch_size=cfg_unet['batchsize'],
               lr=cfg_unet['lr'],
-              run_distribute=args.run_distribute,
               cfg=cfg_unet)