python 代码实现反向传播算法

最新推荐文章于 2023-06-08 14:20:38 发布

DarrenXf

最新推荐文章于 2023-06-08 14:20:38 发布

阅读量1.8k

点赞数 1

分类专栏： python 神经网络深度学习人工智能文章标签：人工智能神经网络

本文链接：https://blog.csdn.net/DarrenXf/article/details/85323359

版权

python 同时被 3 个专栏收录

113 篇文章 1 订阅

订阅专栏

人工智能

41 篇文章 0 订阅

订阅专栏

深度学习

26 篇文章 0 订阅

订阅专栏

实现反向传播的算法可以在python2以及python3中运行,在我的测试环境中可以运行.我并没有详细去测试每一个python版本.
算法中使用的数据是mnist数据集.

下面是算法的代码
forward_neural_network.py

#!/usr/bin/python
# -*- coding: utf-8 -*-
#####################################
# File name : forward_neural_network.py
# Create date : 2018-12-25 20:04
# Modified date : 2018-12-28 14:05
# Author : DARREN
# Describe : not set
# Email : lzygzh@126.com
#####################################
from __future__ import division
from __future__ import print_function

import numpy as np
import mnist_dataset
import matplotlib.pyplot as plt

class DeepModel(object):
    def __init__(self):
        super(DeepModel, self).__init__()
        self.hyper = self._get_hyperparameters()

    def _get_hyperparameters(self):
        dic = {}
        dic["batch_size"] = 128
        dic["epsilon"] = 0.0000001
        dic["reg_lambda"] = 0.05
        dic["learn_rate"] = 0.001
        dic["max_steps"] = 500
        dic["train_steps"] = 1
        dic["max_epochs"] = 20
        dic["input_dim"] = 28*28 #img size
        dic["hidden_dim"] = 2048
        dic["output_dim"] = 10
        dic["file_path"] = "./data/"
        return dic

    def show_hyperparameters(self):
        print("pyperparameters:")
        for key in self.hyper:
            print("%s:%s" % (key, self.hyper[key]))

    def _sig(self, x):
        y = 1 / (1 + np.exp(-x))
        return y

    def _relu(self, x):
        return (np.abs(x) + x) / 2.0

    def _relu_derivative(self, x):
        x[x <= 0] = 0
        x[x > 0] = 1
        return x

    def _sig_deirvative(self, x):
        return self._sig(x) * (1 - self._sig(x))

    def _deal_softmax_numerical(self, logits):
        logits_max = np.max(logits, axis=1)
        for i in range(len(logits)):
            logits[i] = logits[i] - logits_max[i]
        logits = logits + self.hyper["epsilon"]
        return logits

    def _deal_log_numerical(self, probs):
        probs = probs + self.hyper["epsilon"]
        return probs

    def _weight_decay(self, dW2, dW1, W2, W1):
        dW2 += self.hyper["reg_lambda"] * W2
        dW1 += self.hyper["reg_lambda"] * W1
        return dW1, dW2

    def _init_model(self, model=None):
        if model:
            W1 = model["W1"]
            b1 = model["b1"]
            W2 = model["W2"]
            b2 = model["b2"]
        else:
            np.random.seed(0)
            W1 = np.random.randn(self.hyper["input_dim"], self.hyper["hidden_dim"])
            b1 = np.ones((1, self.hyper["hidden_dim"]))
            W2 = np.random.randn(self.hyper["hidden_dim"], self.hyper["output_dim"])
            b2 = np.ones((1, self.hyper["output_dim"]))

            model = {}
            model["W1"] = W1
            model["b1"] = b1
            model["W2"] = W2
            model["b2"] = b2
            model["steps"] = 1
            model["epochs"] = 0
            model["hyperparameters"] = self.hyper
            model["record"] = {}
        return model, W1, b1, W2, b2

    def _normalization(self, batch_img):
        return batch_img / 255.

    def _get_data(self, dataset, data_generator):
        batch_img, batch_labels, status = dataset.get_a_batch_data(data_generator)
        X = batch_img
        X = self._normalization(X)
        Y = batch_labels
        return X, Y, status

    def _print_train_status(self, model):
        print("epoch:%s steps:%s Train_Loss:%2.5f Train_Acc:%2.5f" % (model["epochs"], model["steps"], model["train_loss"], model["train_accuracy"]))

    def _print_test_status(self, model):
        print("epoch:%s steps:%s Train_Loss:%2.5f Test_Loss:%2.5f Train_Acc:%2.5f Test_Acc:%2.5f train_test_gap:%2.5f" % (model["epochs"], model["steps"], model["train_loss"], model["test_loss"], model["train_accuracy"], model["test_accuracy"], model["train_test_gap"]))

    def _forward_propagation(self, model, X, Y):
        model, W1, b1, W2, b2 = self._init_model(model)

        Z1 = np.dot(X, W1)+b1
        a1 = self._relu(Z1)
        #a1 = self._sig(Z1)
        logits = np.dot(a1, W2)+b2
        logits = self._deal_softmax_numerical(logits)
        exp_score = np.exp(logits)
        prob = exp_score/np.sum(exp_score, axis=1, keepdims=1)

        correct_probs = prob[range(X.shape[0]), np.argmax(Y, axis=1)]
        correct_probs = self._deal_log_numerical(correct_probs)
        correct_logprobs = -np.log(correct_probs)

        data_loss = np.sum(correct_logprobs)
        loss = 1./X.shape[0] * data_loss

        pre_Y = np.argmax(prob, axis=1)
        comp = pre_Y == np.argmax(Y, axis=1)
        accuracy = len(np.flatnonzero(comp))/Y.shape[0]

        return model, prob, a1, Z1, loss, accuracy, comp

    def _backward_propagation(self, model, prob, X, Y, a1, Z1):
        W1 = model["W1"]
        W2 = model["W2"]
        dY_pred = prob - Y
        dW2 = np.dot(a1.T, dY_pred)
        da1 = np.dot(dY_pred, W2.T)
        dadZ = self._relu_derivative(Z1)
        #dadZ = self._sig_deirvative(Z1)
        dZ1 = da1 * dadZ
        dW1 = np.dot(X.T, dZ1)
        #dW1,dW2 = self.weight_decay(dW2,dW1,W2,W1)
        model["W2"] += -self.hyper["learn_rate"]*dW2
        model["W1"] += -self.hyper["learn_rate"]*dW1

        return model

    def _core_graph(self, model, X, Y):
        model, prob, a1, Z1, loss, accuracy, comp = self._forward_propagation(model, X, Y)
        model["train_loss"] = loss
        model["train_accuracy"] = accuracy
        model = self._backward_propagation(model, prob, X, Y, a1, Z1)
        return model

    def _train_model_with_epochs(self, model=None):
        dataset = mnist_dataset.MnistSet(self.hyper["file_path"])
        data_generator = dataset.get_train_data_generator(self.hyper["batch_size"])
        while 1:
            X, Y, status = self._get_data(dataset, data_generator)
            if status == False:
                model['epochs'] += 1
                break
            model = self._core_graph(model, X, Y)
            model["steps"] += 1
            if model["steps"] % self.hyper["train_steps"] == 0:
                self._print_train_status(model)

        return model

    def _train_model_with_steps(self, model=None, data_generator=None):
        dataset = mnist_dataset.MnistSet(self.hyper["file_path"])
        if data_generator == None:
            data_generator = dataset.get_train_data_generator(self.hyper["batch_size"])
        while 1:
            X, Y, status = self._get_data(dataset, data_generator)
            if status == False:
                data_generator = dataset.get_train_data_generator(self.hyper["batch_size"])
                model['epochs'] += 1
            model = self._core_graph(model, X, Y)

            model["steps"] += 1
            if model["steps"] % self.hyper["train_steps"] == 0:
                break

        return model, data_generator

    def _test_update_model(self, model, avg_loss, accuracy):
        model["test_loss"] = avg_loss
        model["test_accuracy"] = accuracy
        model["train_test_gap"] = model["train_accuracy"] - model["test_accuracy"]
        return model

    def _test_model(self, model):
        dataset = mnist_dataset.MnistSet(self.hyper["file_path"])
        data_generator = dataset.get_test_data_generator(self.hyper["batch_size"])
        count = 1
        all_correct_numbers = 0
        all_loss = 0.0

        while count:
            X, Y, status = self._get_data(dataset, data_generator)
            if status == False:
                break
            model, prob, a1, Z1, loss, accuracy, comp = self._forward_propagation(model, X, Y)
            all_loss += loss
            all_correct_numbers += len(np.flatnonzero(comp))
            count += 1

        avg_loss = all_loss / count
        accuracy = all_correct_numbers / 10000.0
        self._test_update_model(model, avg_loss, accuracy)
        self._print_test_status(model)
        self._record_model_status(model)
        return model

    def _record_model_status(self, model):
        steps_dic = {}
        steps_dic["epochs"] = model["epochs"]
        steps_dic["steps"] = model["steps"]
        steps_dic["train_loss"] = model["train_loss"]
        steps_dic["train_accuracy"] = model["train_accuracy"]
        steps_dic["test_loss"] = model["test_loss"]
        steps_dic["test_accuracy"] = model["test_accuracy"]
        steps_dic["train_test_gap"] = model["train_test_gap"]
        record = model["record"]
        record[model["steps"]] = steps_dic

    def _plot_record(self, model):
        self._plot_a_key(model, "train_loss", "test_loss")
        self._plot_a_key(model, "train_accuracy", "test_accuracy")

    def _plot_a_key(self, model, train_key, test_key):
        record = model["record"]
        train = []
        test = []
        steps = []
        for key in record:
            steps.append([key])
        steps.sort()
        for i in range(len(steps)):
            step_dic = record[steps[i][0]]
            train_value = step_dic[train_key]
            train.append(train_value)
            test_value = step_dic[test_key]
            test.append(test_value)
        train = np.array(train)
        steps = np.array(steps)
        plt.plot(steps, train)
        plt.plot(steps, test)
        plt.show()

    def run_with_epoch(self):
        model = None
        while 1:
            model = self._train_model_with_epochs(model)
            self._test_model(model)
            if model["epochs"] > self.hyper["max_epochs"]:
                break
        self._plot_record(model)

    def run_with_steps(self):
        model = None
        data_generator = None
        while 1:
            model, data_generator = self._train_model_with_steps(model, data_generator)
            model = self._test_model(model)
            if model["steps"] > self.hyper["max_steps"]:
                break
        self._plot_record(model)

main.py

#!/usr/bin/python
# -*- coding: utf-8 -*-
#####################################
# File name : main.py
# Create date : 2018-12-23 16:53
# Modified date : 2018-12-28 14:05
# Author : DARREN
# Describe : not set
# Email : lzygzh@126.com
#####################################
from __future__ import division
from __future__ import print_function

import forward_neural_network

def test_deep_model_with_epochs():
    neural_model = forward_neural_network.DeepModel()
    neural_model.show_hyperparameters()
    neural_model.run_with_epoch()

def test_deep_model_with_steps():
    neural_model = forward_neural_network.DeepModel()
    neural_model.show_hyperparameters()
    neural_model.run_with_steps()

def run():
    #test_deep_model_with_epochs()
    test_deep_model_with_steps()
    
run()

运行输出
loss:
在这里插入图片描述
accuracy:

pyperparameters:
learn_rate:0.001
epsilon:1e-07
output_dim:10
batch_size:128
file_path:./data/
reg_lambda:0.05
hidden_dim:2048
input_dim:784
max_epochs:20
max_steps:500
train_steps:1
epoch:0 steps:2 Train_Loss:14.85914 Test_Loss:14.04424 Train_Acc:0.07812 Test_Acc:0.11550 train_test_gap:-0.03738
epoch:0 steps:3 Train_Loss:14.10455 Test_Loss:11.44226 Train_Acc:0.12500 Test_Acc:0.27740 train_test_gap:-0.15240
epoch:0 steps:4 Train_Loss:10.32571 Test_Loss:11.86998 Train_Acc:0.35938 Test_Acc:0.25080 train_test_gap:0.10857
epoch:0 steps:5 Train_Loss:12.12818 Test_Loss:11.77254 Train_Acc:0.23438 Test_Acc:0.25760 train_test_gap:-0.02322
epoch:0 steps:6 Train_Loss:11.83764 Test_Loss:10.56273 Train_Acc:0.26562 Test_Acc:0.33440 train_test_gap:-0.06877
epoch:0 steps:7 Train_Loss:11.20711 Test_Loss:9.55890 Train_Acc:0.30469 Test_Acc:0.39460 train_test_gap:-0.08991
epoch:0 steps:8 Train_Loss:9.13490 Test_Loss:5.98574 Train_Acc:0.42969 Test_Acc:0.61920 train_test_gap:-0.18951
epoch:0 steps:9 Train_Loss:6.92577 Test_Loss:6.90012 Train_Acc:0.57031 Test_Acc:0.56260 train_test_gap:0.00771
epoch:0 steps:10 Train_Loss:8.69383 Test_Loss:6.11299 Train_Acc:0.45312 Test_Acc:0.61240 train_test_gap:-0.15928
epoch:0 steps:11 Train_Loss:5.91836 Test_Loss:4.59948 Train_Acc:0.63281 Test_Acc:0.70490 train_test_gap:-0.07209
epoch:0 steps:12 Train_Loss:4.61528 Test_Loss:3.71887 Train_Acc:0.71094 Test_Acc:0.76080 train_test_gap:-0.04986
epoch:0 steps:13 Train_Loss:3.39991 Test_Loss:3.05465 Train_Acc:0.78906 Test_Acc:0.80270 train_test_gap:-0.01364
epoch:0 steps:14 Train_Loss:2.85323 Test_Loss:2.84665 Train_Acc:0.82031 Test_Acc:0.81570 train_test_gap:0.00461
epoch:0 steps:15 Train_Loss:2.51825 Test_Loss:2.80207 Train_Acc:0.84375 Test_Acc:0.81810 train_test_gap:0.02565
epoch:0 steps:16 Train_Loss:2.38831 Test_Loss:2.54680 Train_Acc:0.84375 Test_Acc:0.83350 train_test_gap:0.01025
epoch:0 steps:17 Train_Loss:2.85811 Test_Loss:2.74483 Train_Acc:0.82031 Test_Acc:0.82120 train_test_gap:-0.00089
epoch:0 steps:18 Train_Loss:1.66736 Test_Loss:2.43784 Train_Acc:0.89062 Test_Acc:0.84150 train_test_gap:0.04912
epoch:0 steps:19 Train_Loss:2.51441 Test_Loss:3.04179 Train_Acc:0.83594 Test_Acc:0.80270 train_test_gap:0.03324
epoch:0 steps:20 Train_Loss:3.04864 Test_Loss:2.36385 Train_Acc:0.79688 Test_Acc:0.84590 train_test_gap:-0.04902
epoch:0 steps:21 Train_Loss:2.15603 Test_Loss:3.48260 Train_Acc:0.85156 Test_Acc:0.77610 train_test_gap:0.07546
epoch:0 steps:22 Train_Loss:2.15149 Test_Loss:2.26198 Train_Acc:0.85938 Test_Acc:0.85330 train_test_gap:0.00608
epoch:0 steps:23 Train_Loss:3.26218 Test_Loss:2.66986 Train_Acc:0.78906 Test_Acc:0.82600 train_test_gap:-0.03694
epoch:0 steps:24 Train_Loss:2.51811 Test_Loss:2.53763 Train_Acc:0.84375 Test_Acc:0.83470 train_test_gap:0.00905
epoch:0 steps:25 Train_Loss:2.60409 Test_Loss:2.54477 Train_Acc:0.83594 Test_Acc:0.83410 train_test_gap:0.00184
epoch:0 steps:26 Train_Loss:1.75691 Test_Loss:2.35574 Train_Acc:0.89062 Test_Acc:0.84540 train_test_gap:0.04522
epoch:0 steps:27 Train_Loss:2.13802 Test_Loss:2.10796 Train_Acc:0.86719 Test_Acc:0.86150 train_test_gap:0.00569
epoch:0 steps:28 Train_Loss:1.63699 Test_Loss:2.31571 Train_Acc:0.89844 Test_Acc:0.84780 train_test_gap:0.05064
epoch:0 steps:29 Train_Loss:2.24861 Test_Loss:2.22747 Train_Acc:0.85938 Test_Acc:0.85410 train_test_gap:0.00528
epoch:0 steps:30 Train_Loss:2.01263 Test_Loss:2.08544 Train_Acc:0.87500 Test_Acc:0.86270 train_test_gap:0.01230
epoch:0 steps:31 Train_Loss:2.36653 Test_Loss:1.97203 Train_Acc:0.85156 Test_Acc:0.87020 train_test_gap:-0.01864
epoch:0 steps:32 Train_Loss:1.07500 Test_Loss:2.03386 Train_Acc:0.92188 Test_Acc:0.86650 train_test_gap:0.05537
epoch:0 steps:33 Train_Loss:2.07486 Test_Loss:2.08461 Train_Acc:0.86719 Test_Acc:0.86320 train_test_gap:0.00399
epoch:0 steps:34 Train_Loss:1.78800 Test_Loss:2.15912 Train_Acc:0.88281 Test_Acc:0.85780 train_test_gap:0.02501
epoch:0 steps:35 Train_Loss:2.34563 Test_Loss:2.55244 Train_Acc:0.85156 Test_Acc:0.83330 train_test_gap:0.01826
epoch:0 steps:36 Train_Loss:2.63291 Test_Loss:2.34145 Train_Acc:0.83594 Test_Acc:0.84670 train_test_gap:-0.01076
epoch:0 steps:37 Train_Loss:2.14546 Test_Loss:2.97026 Train_Acc:0.85938 Test_Acc:0.80750 train_test_gap:0.05188
epoch:0 steps:38 Train_Loss:3.27297 Test_Loss:2.43775 Train_Acc:0.79688 Test_Acc:0.84080 train_test_gap:-0.04392
epoch:0 steps:39 Train_Loss:2.02468 Test_Loss:2.09951 Train_Acc:0.86719 Test_Acc:0.86160 train_test_gap:0.00559
epoch:0 steps:40 Train_Loss:2.01476 Test_Loss:2.13885 Train_Acc:0.87500 Test_Acc:0.85940 train_test_gap:0.01560
epoch:0 steps:41 Train_Loss:1.68225 Test_Loss:1.91793 Train_Acc:0.88281 Test_Acc:0.87400 train_test_gap:0.00881
epoch:0 steps:42 Train_Loss:2.32264 Test_Loss:2.29469 Train_Acc:0.85156 Test_Acc:0.84990 train_test_gap:0.00166
epoch:0 steps:43 Train_Loss:2.22248 Test_Loss:2.22907 Train_Acc:0.85156 Test_Acc:0.85460 train_test_gap:-0.00304
epoch:0 steps:44 Train_Loss:2.05484 Test_Loss:1.90333 Train_Acc:0.86719 Test_Acc:0.87430 train_test_gap:-0.00711
epoch:0 steps:45 Train_Loss:2.22903 Test_Loss:1.89846 Train_Acc:0.85156 Test_Acc:0.87590 train_test_gap:-0.02434
epoch:0 steps:46 Train_Loss:1.51116 Test_Loss:1.86426 Train_Acc:0.90625 Test_Acc:0.87650 train_test_gap:0.02975
epoch:0 steps:47 Train_Loss:2.49274 Test_Loss:2.44396 Train_Acc:0.83594 Test_Acc:0.84170 train_test_gap:-0.00576
epoch:0 steps:48 Train_Loss:2.18290 Test_Loss:2.14242 Train_Acc:0.85938 Test_Acc:0.86000 train_test_gap:-0.00062
epoch:0 steps:49 Train_Loss:1.53958 Test_Loss:1.76525 Train_Acc:0.89844 Test_Acc:0.88400 train_test_gap:0.01444
epoch:0 steps:50 Train_Loss:0.75554 Test_Loss:1.73318 Train_Acc:0.95312 Test_Acc:0.88560 train_test_gap:0.06752
epoch:0 steps:51 Train_Loss:1.38258 Test_Loss:2.07076 Train_Acc:0.91406 Test_Acc:0.86380 train_test_gap:0.05026
epoch:0 steps:52 Train_Loss:1.78104 Test_Loss:2.03037 Train_Acc:0.88281 Test_Acc:0.86650 train_test_gap:0.01631
epoch:0 steps:53 Train_Loss:1.16752 Test_Loss:1.74383 Train_Acc:0.92188 Test_Acc:0.88490 train_test_gap:0.03697
epoch:0 steps:54 Train_Loss:1.18224 Test_Loss:1.89805 Train_Acc:0.92188 Test_Acc:0.87480 train_test_gap:0.04707
epoch:0 steps:55 Train_Loss:2.37683 Test_Loss:1.79195 Train_Acc:0.84375 Test_Acc:0.88260 train_test_gap:-0.03885
epoch:0 steps:56 Train_Loss:2.19575 Test_Loss:2.29255 Train_Acc:0.85938 Test_Acc:0.85090 train_test_gap:0.00848
epoch:0 steps:57 Train_Loss:1.84066 Test_Loss:1.84066 Train_Acc:0.86719 Test_Acc:0.87830 train_test_gap:-0.01111
epoch:0 steps:58 Train_Loss:3.08262 Test_Loss:1.87490 Train_Acc:0.80469 Test_Acc:0.87700 train_test_gap:-0.07231
epoch:0 steps:59 Train_Loss:1.68296 Test_Loss:1.85690 Train_Acc:0.89062 Test_Acc:0.87710 train_test_gap:0.01353
epoch:0 steps:60 Train_Loss:1.51256 Test_Loss:1.79702 Train_Acc:0.90625 Test_Acc:0.88150 train_test_gap:0.02475
epoch:0 steps:61 Train_Loss:1.46974 Test_Loss:1.75727 Train_Acc:0.90625 Test_Acc:0.88490 train_test_gap:0.02135
epoch:0 steps:62 Train_Loss:1.77350 Test_Loss:2.14238 Train_Acc:0.88281 Test_Acc:0.86000 train_test_gap:0.02281
epoch:0 steps:63 Train_Loss:3.14911 Test_Loss:2.80807 Train_Acc:0.80469 Test_Acc:0.81750 train_test_gap:-0.01281
epoch:0 steps:64 Train_Loss:3.11754 Test_Loss:2.46263 Train_Acc:0.79688 Test_Acc:0.83910 train_test_gap:-0.04222
epoch:0 steps:65 Train_Loss:2.52628 Test_Loss:1.84501 Train_Acc:0.83594 Test_Acc:0.87760 train_test_gap:-0.04166
epoch:0 steps:66 Train_Loss:2.77292 Test_Loss:1.92371 Train_Acc:0.81250 Test_Acc:0.87410 train_test_gap:-0.06160
epoch:0 steps:67 Train_Loss:1.91380 Test_Loss:1.67202 Train_Acc:0.87500 Test_Acc:0.88880 train_test_gap:-0.01380
epoch:0 steps:68 Train_Loss:1.88037 Test_Loss:1.95678 Train_Acc:0.88281 Test_Acc:0.87070 train_test_gap:0.01211
epoch:0 steps:69 Train_Loss:2.55928 Test_Loss:2.13312 Train_Acc:0.82812 Test_Acc:0.85950 train_test_gap:-0.03138
epoch:0 steps:70 Train_Loss:3.48630 Test_Loss:2.35307 Train_Acc:0.77344 Test_Acc:0.84610 train_test_gap:-0.07266
epoch:0 steps:71 Train_Loss:3.72949 Test_Loss:2.05432 Train_Acc:0.76562 Test_Acc:0.86440 train_test_gap:-0.09877
epoch:0 steps:72 Train_Loss:1.36932 Test_Loss:1.66397 Train_Acc:0.90625 Test_Acc:0.88930 train_test_gap:0.01695
epoch:0 steps:73 Train_Loss:1.76591 Test_Loss:1.60917 Train_Acc:0.87500 Test_Acc:0.89220 train_test_gap:-0.01720
epoch:0 steps:74 Train_Loss:1.84427 Test_Loss:1.65102 Train_Acc:0.87500 Test_Acc:0.89060 train_test_gap:-0.01560
epoch:0 steps:75 Train_Loss:2.04340 Test_Loss:2.15771 Train_Acc:0.86719 Test_Acc:0.85820 train_test_gap:0.00899
epoch:0 steps:76 Train_Loss:2.93961 Test_Loss:2.27221 Train_Acc:0.80469 Test_Acc:0.85130 train_test_gap:-0.04661
epoch:0 steps:77 Train_Loss:1.49200 Test_Loss:1.90459 Train_Acc:0.90625 Test_Acc:0.87390 train_test_gap:0.03235
epoch:0 steps:78 Train_Loss:2.14052 Test_Loss:1.79451 Train_Acc:0.86719 Test_Acc:0.88050 train_test_gap:-0.01331
epoch:0 steps:79 Train_Loss:1.64249 Test_Loss:1.56446 Train_Acc:0.89062 Test_Acc:0.89620 train_test_gap:-0.00557
epoch:0 steps:80 Train_Loss:1.63978 Test_Loss:1.76226 Train_Acc:0.89844 Test_Acc:0.88420 train_test_gap:0.01424
epoch:0 steps:81 Train_Loss:2.22750 Test_Loss:2.01901 Train_Acc:0.85156 Test_Acc:0.86650 train_test_gap:-0.01494
epoch:0 steps:82 Train_Loss:1.20130 Test_Loss:1.91067 Train_Acc:0.91406 Test_Acc:0.87370 train_test_gap:0.04036
epoch:0 steps:83 Train_Loss:1.38000 Test_Loss:1.60397 Train_Acc:0.91406 Test_Acc:0.89420 train_test_gap:0.01986
epoch:0 steps:84 Train_Loss:0.63019 Test_Loss:1.61411 Train_Acc:0.96094 Test_Acc:0.89280 train_test_gap:0.06814
epoch:0 steps:85 Train_Loss:1.18145 Test_Loss:1.52724 Train_Acc:0.92188 Test_Acc:0.89900 train_test_gap:0.02287
epoch:0 steps:86 Train_Loss:1.54103 Test_Loss:1.63063 Train_Acc:0.89844 Test_Acc:0.89100 train_test_gap:0.00744
epoch:0 steps:87 Train_Loss:1.27609 Test_Loss:1.64715 Train_Acc:0.91406 Test_Acc:0.89030 train_test_gap:0.02376
epoch:0 steps:88 Train_Loss:1.20739 Test_Loss:1.73880 Train_Acc:0.92188 Test_Acc:0.88510 train_test_gap:0.03678
epoch:0 steps:89 Train_Loss:1.51107 Test_Loss:1.60485 Train_Acc:0.90625 Test_Acc:0.89360 train_test_gap:0.01265
epoch:0 steps:90 Train_Loss:1.54020 Test_Loss:1.69655 Train_Acc:0.89844 Test_Acc:0.88770 train_test_gap:0.01074
epoch:0 steps:91 Train_Loss:1.28825 Test_Loss:1.67608 Train_Acc:0.91406 Test_Acc:0.88890 train_test_gap:0.02516
epoch:0 steps:92 Train_Loss:2.09532 Test_Loss:2.04284 Train_Acc:0.85938 Test_Acc:0.86420 train_test_gap:-0.00482
epoch:0 steps:93 Train_Loss:2.00806 Test_Loss:1.99674 Train_Acc:0.87500 Test_Acc:0.86720 train_test_gap:0.00780
epoch:0 steps:94 Train_Loss:2.56012 Test_Loss:1.50437 Train_Acc:0.83594 Test_Acc:0.90030 train_test_gap:-0.06436
epoch:0 steps:95 Train_Loss:1.51107 Test_Loss:1.60177 Train_Acc:0.90625 Test_Acc:0.89430 train_test_gap:0.01195
epoch:0 steps:96 Train_Loss:1.25034 Test_Loss:1.69078 Train_Acc:0.92188 Test_Acc:0.88850 train_test_gap:0.03338
epoch:0 steps:97 Train_Loss:1.49119 Test_Loss:1.67512 Train_Acc:0.89844 Test_Acc:0.88830 train_test_gap:0.01014
epoch:0 steps:98 Train_Loss:1.29241 Test_Loss:1.53756 Train_Acc:0.91406 Test_Acc:0.89790 train_test_gap:0.01616
epoch:0 steps:99 Train_Loss:2.18005 Test_Loss:1.72102 Train_Acc:0.85938 Test_Acc:0.88580 train_test_gap:-0.02643
epoch:0 steps:100 Train_Loss:2.74346 Test_Loss:1.52496 Train_Acc:0.82031 Test_Acc:0.89860 train_test_gap:-0.07829
epoch:0 steps:101 Train_Loss:1.51191 Test_Loss:1.57579 Train_Acc:0.90625 Test_Acc:0.89450 train_test_gap:0.01175
epoch:0 steps:102 Train_Loss:1.84709 Test_Loss:1.70078 Train_Acc:0.87500 Test_Acc:0.88810 train_test_gap:-0.01310
epoch:0 steps:103 Train_Loss:1.96365 Test_Loss:1.61565 Train_Acc:0.86719 Test_Acc:0.89250 train_test_gap:-0.02531
epoch:0 steps:104 Train_Loss:2.14068 Test_Loss:1.47719 Train_Acc:0.86719 Test_Acc:0.90120 train_test_gap:-0.03401
epoch:0 steps:105 Train_Loss:1.13120 Test_Loss:1.49931 Train_Acc:0.92969 Test_Acc:0.90010 train_test_gap:0.02959
epoch:0 steps:106 Train_Loss:1.76337 Test_Loss:1.53826 Train_Acc:0.89062 Test_Acc:0.89780 train_test_gap:-0.00718
epoch:0 steps:107 Train_Loss:1.81708 Test_Loss:1.48611 Train_Acc:0.88281 Test_Acc:0.90200 train_test_gap:-0.01919
epoch:0 steps:108 Train_Loss:1.03029 Test_Loss:1.46395 Train_Acc:0.92969 Test_Acc:0.90250 train_test_gap:0.02719
epoch:0 steps:109 Train_Loss:2.04661 Test_Loss:1.48251 Train_Acc:0.85938 Test_Acc:0.90080 train_test_gap:-0.04143
epoch:0 steps:110 Train_Loss:1.83475 Test_Loss:1.47596 Train_Acc:0.87500 Test_Acc:0.90080 train_test_gap:-0.02580
epoch:0 steps:111 Train_Loss:1.85952 Test_Loss:1.42591 Train_Acc:0.87500 Test_Acc:0.90540 train_test_gap:-0.03040
epoch:0 steps:112 Train_Loss:1.88752 Test_Loss:1.50996 Train_Acc:0.88281 Test_Acc:0.89910 train_test_gap:-0.01629
epoch:0 steps:113 Train_Loss:1.85374 Test_Loss:1.52481 Train_Acc:0.88281 Test_Acc:0.89790 train_test_gap:-0.01509
epoch:0 steps:114 Train_Loss:1.69140 Test_Loss:1.51936 Train_Acc:0.89062 Test_Acc:0.89930 train_test_gap:-0.00867
epoch:0 steps:115 Train_Loss:1.65337 Test_Loss:1.71504 Train_Acc:0.89062 Test_Acc:0.88620 train_test_gap:0.00443
epoch:0 steps:116 Train_Loss:2.01460 Test_Loss:2.17072 Train_Acc:0.87500 Test_Acc:0.85600 train_test_gap:0.01900
epoch:0 steps:117 Train_Loss:4.08097 Test_Loss:2.92189 Train_Acc:0.74219 Test_Acc:0.80870 train_test_gap:-0.06651
epoch:0 steps:118 Train_Loss:2.23922 Test_Loss:1.51141 Train_Acc:0.83594 Test_Acc:0.89870 train_test_gap:-0.06276
epoch:0 steps:119 Train_Loss:0.75554 Test_Loss:1.38668 Train_Acc:0.95312 Test_Acc:0.90670 train_test_gap:0.04643
epoch:0 steps:120 Train_Loss:2.19887 Test_Loss:1.43029 Train_Acc:0.85938 Test_Acc:0.90400 train_test_gap:-0.04463
epoch:0 steps:121 Train_Loss:2.51845 Test_Loss:1.56857 Train_Acc:0.84375 Test_Acc:0.89620 train_test_gap:-0.05245
epoch:0 steps:122 Train_Loss:1.31535 Test_Loss:1.32084 Train_Acc:0.90625 Test_Acc:0.91180 train_test_gap:-0.00555
epoch:0 steps:123 Train_Loss:1.72758 Test_Loss:1.51842 Train_Acc:0.89062 Test_Acc:0.89840 train_test_gap:-0.00777
epoch:0 steps:124 Train_Loss:1.26191 Test_Loss:1.37800 Train_Acc:0.92188 Test_Acc:0.90800 train_test_gap:0.01387
epoch:0 steps:125 Train_Loss:2.09344 Test_Loss:1.33070 Train_Acc:0.85938 Test_Acc:0.91180 train_test_gap:-0.05243
epoch:0 steps:126 Train_Loss:1.67811 Test_Loss:1.44686 Train_Acc:0.89062 Test_Acc:0.90310 train_test_gap:-0.01248
epoch:0 steps:127 Train_Loss:2.13772 Test_Loss:1.34152 Train_Acc:0.85938 Test_Acc:0.91040 train_test_gap:-0.05102
epoch:0 steps:128 Train_Loss:1.20256 Test_Loss:1.53400 Train_Acc:0.92188 Test_Acc:0.89660 train_test_gap:0.02528
epoch:0 steps:129 Train_Loss:1.58165 Test_Loss:1.31991 Train_Acc:0.89062 Test_Acc:0.91130 train_test_gap:-0.02067
epoch:0 steps:130 Train_Loss:1.11436 Test_Loss:1.45791 Train_Acc:0.92969 Test_Acc:0.90270 train_test_gap:0.02699
epoch:0 steps:131 Train_Loss:1.25922 Test_Loss:1.44754 Train_Acc:0.92188 Test_Acc:0.90320 train_test_gap:0.01867
epoch:0 steps:132 Train_Loss:2.00345 Test_Loss:1.35603 Train_Acc:0.87500 Test_Acc:0.90870 train_test_gap:-0.03370
epoch:0 steps:133 Train_Loss:1.49630 Test_Loss:1.26369 Train_Acc:0.89844 Test_Acc:0.91480 train_test_gap:-0.01636
epoch:0 steps:134 Train_Loss:2.14140 Test_Loss:1.39592 Train_Acc:0.86719 Test_Acc:0.90700 train_test_gap:-0.03981
epoch:0 steps:135 Train_Loss:2.13496 Test_Loss:1.51698 Train_Acc:0.86719 Test_Acc:0.89810 train_test_gap:-0.03091
epoch:0 steps:136 Train_Loss:1.67060 Test_Loss:1.33204 Train_Acc:0.89062 Test_Acc:0.90990 train_test_gap:-0.01928
epoch:0 steps:137 Train_Loss:1.25676 Test_Loss:1.34804 Train_Acc:0.92188 Test_Acc:0.91130 train_test_gap:0.01058
epoch:0 steps:138 Train_Loss:2.39663 Test_Loss:1.79700 Train_Acc:0.84375 Test_Acc:0.87960 train_test_gap:-0.03585
epoch:0 steps:139 Train_Loss:2.26671 Test_Loss:1.64889 Train_Acc:0.85938 Test_Acc:0.88980 train_test_gap:-0.03043
epoch:0 steps:140 Train_Loss:2.15292 Test_Loss:1.51108 Train_Acc:0.85938 Test_Acc:0.89940 train_test_gap:-0.04002
epoch:0 steps:141 Train_Loss:1.88928 Test_Loss:1.76333 Train_Acc:0.88281 Test_Acc:0.88320 train_test_gap:-0.00039
epoch:0 steps:142 Train_Loss:2.54066 Test_Loss:1.91654 Train_Acc:0.83594 Test_Acc:0.87270 train_test_gap:-0.03676
epoch:0 steps:143 Train_Loss:1.38515 Test_Loss:1.28058 Train_Acc:0.91406 Test_Acc:0.91390 train_test_gap:0.00016
epoch:0 steps:144 Train_Loss:0.58681 Test_Loss:1.36649 Train_Acc:0.96094 Test_Acc:0.90790 train_test_gap:0.05304
epoch:0 steps:145 Train_Loss:1.36598 Test_Loss:1.46186 Train_Acc:0.91406 Test_Acc:0.90150 train_test_gap:0.01256
epoch:0 steps:146 Train_Loss:1.08951 Test_Loss:1.77855 Train_Acc:0.91406 Test_Acc:0.88230 train_test_gap:0.03176
epoch:0 steps:147 Train_Loss:1.29163 Test_Loss:1.28077 Train_Acc:0.91406 Test_Acc:0.91390 train_test_gap:0.00016
epoch:0 steps:148 Train_Loss:1.00738 Test_Loss:1.34841 Train_Acc:0.93750 Test_Acc:0.91050 train_test_gap:0.02700
epoch:0 steps:149 Train_Loss:1.24775 Test_Loss:1.44683 Train_Acc:0.92188 Test_Acc:0.90300 train_test_gap:0.01887
epoch:0 steps:150 Train_Loss:1.63699 Test_Loss:1.24404 Train_Acc:0.89844 Test_Acc:0.91630 train_test_gap:-0.01786
epoch:0 steps:151 Train_Loss:1.40818 Test_Loss:1.25434 Train_Acc:0.90625 Test_Acc:0.91640 train_test_gap:-0.01015
epoch:0 steps:152 Train_Loss:0.88322 Test_Loss:1.38043 Train_Acc:0.94531 Test_Acc:0.90750 train_test_gap:0.03781
epoch:0 steps:153 Train_Loss:1.25924 Test_Loss:1.28706 Train_Acc:0.92188 Test_Acc:0.91330 train_test_gap:0.00857
epoch:0 steps:154 Train_Loss:1.12511 Test_Loss:1.47410 Train_Acc:0.92969 Test_Acc:0.90140 train_test_gap:0.02829
epoch:0 steps:155 Train_Loss:0.92301 Test_Loss:1.40405 Train_Acc:0.93750 Test_Acc:0.90580 train_test_gap:0.03170
epoch:0 steps:156 Train_Loss:1.38514 Test_Loss:1.21478 Train_Acc:0.91406 Test_Acc:0.91810 train_test_gap:-0.00404
epoch:0 steps:157 Train_Loss:1.00738 Test_Loss:1.77479 Train_Acc:0.93750 Test_Acc:0.88180 train_test_gap:0.05570
epoch:0 steps:158 Train_Loss:2.70003 Test_Loss:1.92220 Train_Acc:0.82812 Test_Acc:0.87270 train_test_gap:-0.04458
epoch:0 steps:159 Train_Loss:2.26363 Test_Loss:1.63358 Train_Acc:0.84375 Test_Acc:0.89160 train_test_gap:-0.04785
epoch:0 steps:160 Train_Loss:1.15428 Test_Loss:1.49389 Train_Acc:0.91406 Test_Acc:0.90060 train_test_gap:0.01346
epoch:0 steps:161 Train_Loss:1.44622 Test_Loss:1.29615 Train_Acc:0.89844 Test_Acc:0.91400 train_test_gap:-0.01556
epoch:0 steps:162 Train_Loss:1.12965 Test_Loss:1.27349 Train_Acc:0.92188 Test_Acc:0.91360 train_test_gap:0.00828
epoch:0 steps:163 Train_Loss:1.71512 Test_Loss:1.44025 Train_Acc:0.89062 Test_Acc:0.90380 train_test_gap:-0.01318
epoch:0 steps:164 Train_Loss:1.50991 Test_Loss:1.61560 Train_Acc:0.90625 Test_Acc:0.89240 train_test_gap:0.01385
epoch:0 steps:165 Train_Loss:2.46042 Test_Loss:1.72226 Train_Acc:0.84375 Test_Acc:0.88510 train_test_gap:-0.04135
epoch:0 steps:166 Train_Loss:2.22030 Test_Loss:1.25728 Train_Acc:0.85938 Test_Acc:0.91550 train_test_gap:-0.05612
epoch:0 steps:167 Train_Loss:0.82783 Test_Loss:1.18729 Train_Acc:0.94531 Test_Acc:0.91950 train_test_gap:0.02581
epoch:0 steps:168 Train_Loss:1.03260 Test_Loss:1.28701 Train_Acc:0.92188 Test_Acc:0.91340 train_test_gap:0.00848
epoch:0 steps:169 Train_Loss:1.51107 Test_Loss:1.42736 Train_Acc:0.90625 Test_Acc:0.90380 train_test_gap:0.00245
epoch:0 steps:170 Train_Loss:1.38022 Test_Loss:1.32062 Train_Acc:0.90625 Test_Acc:0.91140 train_test_gap:-0.00515
epoch:0 steps:171 Train_Loss:1.25927 Test_Loss:1.25867 Train_Acc:0.92188 Test_Acc:0.91480 train_test_gap:0.00708
epoch:0 steps:172 Train_Loss:0.44530 Test_Loss:1.26718 Train_Acc:0.96875 Test_Acc:0.91520 train_test_gap:0.05355
epoch:0 steps:173 Train_Loss:1.72009 Test_Loss:1.46640 Train_Acc:0.88281 Test_Acc:0.90140 train_test_gap:-0.01859
epoch:0 steps:174 Train_Loss:2.02750 Test_Loss:1.69696 Train_Acc:0.86719 Test_Acc:0.88740 train_test_gap:-0.02021
epoch:0 steps:175 Train_Loss:2.41334 Test_Loss:1.64947 Train_Acc:0.84375 Test_Acc:0.88910 train_test_gap:-0.04535
epoch:0 steps:176 Train_Loss:1.14247 Test_Loss:1.33067 Train_Acc:0.92188 Test_Acc:0.90970 train_test_gap:0.01218
epoch:0 steps:177 Train_Loss:2.57475 Test_Loss:1.38983 Train_Acc:0.82031 Test_Acc:0.90730 train_test_gap:-0.08699
epoch:0 steps:178 Train_Loss:2.17181 Test_Loss:1.51738 Train_Acc:0.85938 Test_Acc:0.89800 train_test_gap:-0.03863
epoch:0 steps:179 Train_Loss:2.68798 Test_Loss:1.29887 Train_Acc:0.81250 Test_Acc:0.91230 train_test_gap:-0.09980
epoch:0 steps:180 Train_Loss:1.40679 Test_Loss:1.25023 Train_Acc:0.90625 Test_Acc:0.91590 train_test_gap:-0.00965
epoch:0 steps:181 Train_Loss:0.75554 Test_Loss:1.30939 Train_Acc:0.95312 Test_Acc:0.91150 train_test_gap:0.04163
epoch:0 steps:182 Train_Loss:0.96149 Test_Loss:1.17130 Train_Acc:0.92969 Test_Acc:0.92130 train_test_gap:0.00839
epoch:0 steps:183 Train_Loss:1.25923 Test_Loss:1.17130 Train_Acc:0.92188 Test_Acc:0.92070 train_test_gap:0.00118
epoch:0 steps:184 Train_Loss:1.21710 Test_Loss:1.18762 Train_Acc:0.92188 Test_Acc:0.92020 train_test_gap:0.00167
epoch:0 steps:185 Train_Loss:1.11595 Test_Loss:1.25638 Train_Acc:0.92188 Test_Acc:0.91560 train_test_gap:0.00628
epoch:0 steps:186 Train_Loss:1.47181 Test_Loss:1.27399 Train_Acc:0.89844 Test_Acc:0.91330 train_test_gap:-0.01486
epoch:0 steps:187 Train_Loss:1.92982 Test_Loss:1.28321 Train_Acc:0.86719 Test_Acc:0.91340 train_test_gap:-0.04621
epoch:0 steps:188 Train_Loss:1.13997 Test_Loss:1.12199 Train_Acc:0.92188 Test_Acc:0.92340 train_test_gap:-0.00152
epoch:0 steps:189 Train_Loss:1.23791 Test_Loss:1.23675 Train_Acc:0.92188 Test_Acc:0.91640 train_test_gap:0.00548
epoch:0 steps:190 Train_Loss:0.50404 Test_Loss:1.17043 Train_Acc:0.96875 Test_Acc:0.92000 train_test_gap:0.04875
epoch:0 steps:191 Train_Loss:1.26661 Test_Loss:1.32719 Train_Acc:0.91406 Test_Acc:0.91010 train_test_gap:0.00396
epoch:0 steps:192 Train_Loss:0.62961 Test_Loss:1.22283 Train_Acc:0.96094 Test_Acc:0.91710 train_test_gap:0.04384
epoch:0 steps:193 Train_Loss:2.01473 Test_Loss:1.26009 Train_Acc:0.87500 Test_Acc:0.91540 train_test_gap:-0.04040
epoch:0 steps:194 Train_Loss:1.76619 Test_Loss:1.18470 Train_Acc:0.88281 Test_Acc:0.91950 train_test_gap:-0.03669
epoch:0 steps:195 Train_Loss:1.74815 Test_Loss:1.49404 Train_Acc:0.89062 Test_Acc:0.89920 train_test_gap:-0.00857
epoch:0 steps:196 Train_Loss:1.39463 Test_Loss:1.19238 Train_Acc:0.89844 Test_Acc:0.91930 train_test_gap:-0.02086
epoch:0 steps:197 Train_Loss:0.88146 Test_Loss:1.15315 Train_Acc:0.94531 Test_Acc:0.92200 train_test_gap:0.02331
epoch:0 steps:198 Train_Loss:1.15282 Test_Loss:1.38388 Train_Acc:0.92188 Test_Acc:0.90630 train_test_gap:0.01558
epoch:0 steps:199 Train_Loss:1.88884 Test_Loss:1.45151 Train_Acc:0.88281 Test_Acc:0.90270 train_test_gap:-0.01989
epoch:0 steps:200 Train_Loss:0.82203 Test_Loss:1.22517 Train_Acc:0.94531 Test_Acc:0.91720 train_test_gap:0.02811
epoch:0 steps:201 Train_Loss:1.25956 Test_Loss:1.20319 Train_Acc:0.92188 Test_Acc:0.91780 train_test_gap:0.00408
epoch:0 steps:202 Train_Loss:1.00628 Test_Loss:1.13779 Train_Acc:0.93750 Test_Acc:0.92280 train_test_gap:0.01470
epoch:0 steps:203 Train_Loss:1.66224 Test_Loss:1.27571 Train_Acc:0.89062 Test_Acc:0.91390 train_test_gap:-0.02328
epoch:0 steps:204 Train_Loss:1.31582 Test_Loss:1.15462 Train_Acc:0.90625 Test_Acc:0.92190 train_test_gap:-0.01565
epoch:0 steps:205 Train_Loss:0.74381 Test_Loss:1.15657 Train_Acc:0.94531 Test_Acc:0.92120 train_test_gap:0.02411
epoch:0 steps:206 Train_Loss:0.62961 Test_Loss:1.21480 Train_Acc:0.96094 Test_Acc:0.91780 train_test_gap:0.04314
epoch:0 steps:207 Train_Loss:1.28379 Test_Loss:1.14159 Train_Acc:0.91406 Test_Acc:0.92290 train_test_gap:-0.00884
epoch:0 steps:208 Train_Loss:2.06020 Test_Loss:1.19667 Train_Acc:0.85938 Test_Acc:0.91950 train_test_gap:-0.06012
epoch:0 steps:209 Train_Loss:1.53985 Test_Loss:1.19692 Train_Acc:0.89844 Test_Acc:0.91920 train_test_gap:-0.02076
epoch:0 steps:210 Train_Loss:1.64194 Test_Loss:1.20831 Train_Acc:0.89844 Test_Acc:0.91940 train_test_gap:-0.02096
epoch:0 steps:211 Train_Loss:1.00820 Test_Loss:1.18922 Train_Acc:0.93750 Test_Acc:0.92010 train_test_gap:0.01740
epoch:0 steps:212 Train_Loss:0.81380 Test_Loss:1.22392 Train_Acc:0.94531 Test_Acc:0.91780 train_test_gap:0.02751
epoch:0 steps:213 Train_Loss:0.63468 Test_Loss:1.23514 Train_Acc:0.96094 Test_Acc:0.91640 train_test_gap:0.04454
epoch:0 steps:214 Train_Loss:2.13671 Test_Loss:1.79122 Train_Acc:0.86719 Test_Acc:0.88020 train_test_gap:-0.01301
epoch:0 steps:215 Train_Loss:1.57693 Test_Loss:1.40291 Train_Acc:0.89062 Test_Acc:0.90600 train_test_gap:-0.01538
epoch:0 steps:216 Train_Loss:1.75634 Test_Loss:1.33256 Train_Acc:0.88281 Test_Acc:0.90940 train_test_gap:-0.02659
epoch:0 steps:217 Train_Loss:0.86231 Test_Loss:1.45306 Train_Acc:0.94531 Test_Acc:0.90240 train_test_gap:0.04291
epoch:0 steps:218 Train_Loss:1.72893 Test_Loss:1.25944 Train_Acc:0.88281 Test_Acc:0.91560 train_test_gap:-0.03279
epoch:0 steps:219 Train_Loss:1.51703 Test_Loss:1.16232 Train_Acc:0.89062 Test_Acc:0.92100 train_test_gap:-0.03038
epoch:0 steps:220 Train_Loss:1.31595 Test_Loss:1.46626 Train_Acc:0.90625 Test_Acc:0.90090 train_test_gap:0.00535
epoch:0 steps:221 Train_Loss:1.38515 Test_Loss:1.22163 Train_Acc:0.91406 Test_Acc:0.91690 train_test_gap:-0.00284
epoch:0 steps:222 Train_Loss:1.02214 Test_Loss:1.26595 Train_Acc:0.92969 Test_Acc:0.91400 train_test_gap:0.01569
epoch:0 steps:223 Train_Loss:1.15895 Test_Loss:1.31244 Train_Acc:0.92188 Test_Acc:0.91170 train_test_gap:0.01018
epoch:0 steps:224 Train_Loss:0.88146 Test_Loss:1.24281 Train_Acc:0.94531 Test_Acc:0.91550 train_test_gap:0.02981
epoch:0 steps:225 Train_Loss:1.51942 Test_Loss:1.38579 Train_Acc:0.90625 Test_Acc:0.90700 train_test_gap:-0.00075
epoch:0 steps:226 Train_Loss:1.09756 Test_Loss:1.41233 Train_Acc:0.92969 Test_Acc:0.90520 train_test_gap:0.02449
epoch:0 steps:227 Train_Loss:1.01027 Test_Loss:1.19937 Train_Acc:0.92188 Test_Acc:0.91870 train_test_gap:0.00318
epoch:0 steps:228 Train_Loss:1.25919 Test_Loss:1.23906 Train_Acc:0.92188 Test_Acc:0.91550 train_test_gap:0.00638
epoch:0 steps:229 Train_Loss:2.56135 Test_Loss:1.63822 Train_Acc:0.83594 Test_Acc:0.89090 train_test_gap:-0.05496
epoch:0 steps:230 Train_Loss:1.51173 Test_Loss:1.20935 Train_Acc:0.90625 Test_Acc:0.91840 train_test_gap:-0.01215
epoch:0 steps:231 Train_Loss:0.87060 Test_Loss:1.56878 Train_Acc:0.93750 Test_Acc:0.89430 train_test_gap:0.04320
epoch:0 steps:232 Train_Loss:1.62506 Test_Loss:1.44184 Train_Acc:0.89062 Test_Acc:0.90310 train_test_gap:-0.01248
epoch:0 steps:233 Train_Loss:0.79490 Test_Loss:1.15310 Train_Acc:0.92969 Test_Acc:0.92200 train_test_gap:0.00769
epoch:0 steps:234 Train_Loss:1.28471 Test_Loss:1.29480 Train_Acc:0.90625 Test_Acc:0.91370 train_test_gap:-0.00745
epoch:0 steps:235 Train_Loss:1.73581 Test_Loss:1.47668 Train_Acc:0.89062 Test_Acc:0.89990 train_test_gap:-0.00928
epoch:0 steps:236 Train_Loss:2.00259 Test_Loss:1.47553 Train_Acc:0.87500 Test_Acc:0.90000 train_test_gap:-0.02500
epoch:0 steps:237 Train_Loss:1.30884 Test_Loss:1.76494 Train_Acc:0.91406 Test_Acc:0.88250 train_test_gap:0.03156
epoch:0 steps:238 Train_Loss:1.13331 Test_Loss:1.26364 Train_Acc:0.92969 Test_Acc:0.91400 train_test_gap:0.01569
epoch:0 steps:239 Train_Loss:0.85171 Test_Loss:1.31986 Train_Acc:0.94531 Test_Acc:0.91090 train_test_gap:0.03441
epoch:0 steps:240 Train_Loss:1.62106 Test_Loss:1.16038 Train_Acc:0.87500 Test_Acc:0.92150 train_test_gap:-0.04650
epoch:0 steps:241 Train_Loss:1.86938 Test_Loss:1.20581 Train_Acc:0.87500 Test_Acc:0.91880 train_test_gap:-0.04380
epoch:0 steps:242 Train_Loss:1.28476 Test_Loss:1.18368 Train_Acc:0.91406 Test_Acc:0.91940 train_test_gap:-0.00534
epoch:0 steps:243 Train_Loss:1.06198 Test_Loss:1.35861 Train_Acc:0.92969 Test_Acc:0.90740 train_test_gap:0.02229
epoch:0 steps:244 Train_Loss:1.86778 Test_Loss:1.27846 Train_Acc:0.87500 Test_Acc:0.91260 train_test_gap:-0.03760
epoch:0 steps:245 Train_Loss:1.25840 Test_Loss:1.20837 Train_Acc:0.92188 Test_Acc:0.91830 train_test_gap:0.00357
epoch:0 steps:246 Train_Loss:2.23984 Test_Loss:1.33443 Train_Acc:0.85938 Test_Acc:0.90970 train_test_gap:-0.05032
epoch:0 steps:247 Train_Loss:1.61115 Test_Loss:1.52094 Train_Acc:0.89844 Test_Acc:0.89720 train_test_gap:0.00124
epoch:0 steps:248 Train_Loss:1.05730 Test_Loss:1.24247 Train_Acc:0.92188 Test_Acc:0.91440 train_test_gap:0.00748
epoch:0 steps:249 Train_Loss:2.26649 Test_Loss:1.22348 Train_Acc:0.85938 Test_Acc:0.91720 train_test_gap:-0.05783
epoch:0 steps:250 Train_Loss:1.32144 Test_Loss:1.53600 Train_Acc:0.91406 Test_Acc:0.89480 train_test_gap:0.01926
epoch:0 steps:251 Train_Loss:1.96968 Test_Loss:1.77802 Train_Acc:0.85938 Test_Acc:0.88050 train_test_gap:-0.02112
epoch:0 steps:252 Train_Loss:2.38680 Test_Loss:1.19642 Train_Acc:0.85156 Test_Acc:0.91970 train_test_gap:-0.06814
epoch:0 steps:253 Train_Loss:1.22697 Test_Loss:1.16363 Train_Acc:0.91406 Test_Acc:0.92100 train_test_gap:-0.00694
epoch:0 steps:254 Train_Loss:1.53530 Test_Loss:1.26815 Train_Acc:0.89844 Test_Acc:0.91490 train_test_gap:-0.01646
epoch:0 steps:255 Train_Loss:1.38571 Test_Loss:1.21016 Train_Acc:0.91406 Test_Acc:0.91800 train_test_gap:-0.00394
epoch:0 steps:256 Train_Loss:0.62961 Test_Loss:1.12885 Train_Acc:0.96094 Test_Acc:0.92350 train_test_gap:0.03744
epoch:0 steps:257 Train_Loss:1.04743 Test_Loss:1.09772 Train_Acc:0.92969 Test_Acc:0.92490 train_test_gap:0.00479
epoch:0 steps:258 Train_Loss:0.85778 Test_Loss:1.11687 Train_Acc:0.94531 Test_Acc:0.92420 train_test_gap:0.02111
epoch:0 steps:259 Train_Loss:1.21742 Test_Loss:1.11771 Train_Acc:0.92188 Test_Acc:0.92410 train_test_gap:-0.00223
epoch:0 steps:260 Train_Loss:0.94319 Test_Loss:1.08788 Train_Acc:0.93750 Test_Acc:0.92590 train_test_gap:0.01160
epoch:0 steps:261 Train_Loss:0.62904 Test_Loss:1.09381 Train_Acc:0.96094 Test_Acc:0.92640 train_test_gap:0.03454
epoch:0 steps:262 Train_Loss:1.49894 Test_Loss:1.07826 Train_Acc:0.90625 Test_Acc:0.92700 train_test_gap:-0.02075
epoch:0 steps:263 Train_Loss:1.51832 Test_Loss:1.07554 Train_Acc:0.89844 Test_Acc:0.92610 train_test_gap:-0.02766
epoch:0 steps:264 Train_Loss:0.82967 Test_Loss:1.17448 Train_Acc:0.94531 Test_Acc:0.91980 train_test_gap:0.02551
epoch:0 steps:265 Train_Loss:0.63541 Test_Loss:1.14211 Train_Acc:0.95312 Test_Acc:0.92260 train_test_gap:0.03053
epoch:0 steps:266 Train_Loss:0.50376 Test_Loss:1.14779 Train_Acc:0.96875 Test_Acc:0.92130 train_test_gap:0.04745
epoch:0 steps:267 Train_Loss:0.25688 Test_Loss:1.18465 Train_Acc:0.98438 Test_Acc:0.91950 train_test_gap:0.06488
epoch:0 steps:268 Train_Loss:0.88149 Test_Loss:1.05350 Train_Acc:0.94531 Test_Acc:0.92810 train_test_gap:0.01721
epoch:0 steps:269 Train_Loss:1.13196 Test_Loss:1.04269 Train_Acc:0.92969 Test_Acc:0.92870 train_test_gap:0.00099
epoch:0 steps:270 Train_Loss:0.84788 Test_Loss:1.10183 Train_Acc:0.93750 Test_Acc:0.92530 train_test_gap:0.01220
epoch:0 steps:271 Train_Loss:1.48261 Test_Loss:1.13723 Train_Acc:0.90625 Test_Acc:0.92160 train_test_gap:-0.01535
epoch:0 steps:272 Train_Loss:1.25979 Test_Loss:1.07195 Train_Acc:0.92188 Test_Acc:0.92620 train_test_gap:-0.00433
epoch:0 steps:273 Train_Loss:1.36419 Test_Loss:1.21945 Train_Acc:0.91406 Test_Acc:0.91750 train_test_gap:-0.00344
epoch:0 steps:274 Train_Loss:1.63207 Test_Loss:1.13563 Train_Acc:0.89844 Test_Acc:0.92190 train_test_gap:-0.02346
epoch:0 steps:275 Train_Loss:0.86407 Test_Loss:1.16149 Train_Acc:0.93750 Test_Acc:0.92030 train_test_gap:0.01720
epoch:0 steps:276 Train_Loss:0.63120 Test_Loss:1.07298 Train_Acc:0.96094 Test_Acc:0.92690 train_test_gap:0.03404
epoch:0 steps:277 Train_Loss:1.49396 Test_Loss:1.38854 Train_Acc:0.90625 Test_Acc:0.90680 train_test_gap:-0.00055
epoch:0 steps:278 Train_Loss:1.26032 Test_Loss:1.22055 Train_Acc:0.92188 Test_Acc:0.91680 train_test_gap:0.00508
epoch:0 steps:279 Train_Loss:1.25923 Test_Loss:1.27749 Train_Acc:0.92188 Test_Acc:0.91430 train_test_gap:0.00757
epoch:0 steps:280 Train_Loss:1.26938 Test_Loss:1.19172 Train_Acc:0.90625 Test_Acc:0.91940 train_test_gap:-0.01315
epoch:0 steps:281 Train_Loss:0.77923 Test_Loss:1.15949 Train_Acc:0.93750 Test_Acc:0.92130 train_test_gap:0.01620
epoch:0 steps:282 Train_Loss:1.36899 Test_Loss:1.08533 Train_Acc:0.90625 Test_Acc:0.92640 train_test_gap:-0.02015
epoch:0 steps:283 Train_Loss:2.07244 Test_Loss:1.25905 Train_Acc:0.86719 Test_Acc:0.91540 train_test_gap:-0.04821
epoch:0 steps:284 Train_Loss:0.88146 Test_Loss:1.20061 Train_Acc:0.94531 Test_Acc:0.91890 train_test_gap:0.02641
epoch:0 steps:285 Train_Loss:0.88146 Test_Loss:1.08207 Train_Acc:0.94531 Test_Acc:0.92640 train_test_gap:0.01891
epoch:0 steps:286 Train_Loss:1.50914 Test_Loss:1.12008 Train_Acc:0.89062 Test_Acc:0.92300 train_test_gap:-0.03238
epoch:0 steps:287 Train_Loss:1.35192 Test_Loss:1.41013 Train_Acc:0.91406 Test_Acc:0.90510 train_test_gap:0.00896
epoch:0 steps:288 Train_Loss:0.50544 Test_Loss:1.29997 Train_Acc:0.96875 Test_Acc:0.91280 train_test_gap:0.05595
epoch:0 steps:289 Train_Loss:1.36528 Test_Loss:1.11911 Train_Acc:0.91406 Test_Acc:0.92380 train_test_gap:-0.00974
epoch:0 steps:290 Train_Loss:0.78691 Test_Loss:1.23311 Train_Acc:0.94531 Test_Acc:0.91640 train_test_gap:0.02891
epoch:0 steps:291 Train_Loss:1.15107 Test_Loss:1.14925 Train_Acc:0.92188 Test_Acc:0.92200 train_test_gap:-0.00013
epoch:0 steps:292 Train_Loss:1.19881 Test_Loss:1.50941 Train_Acc:0.92188 Test_Acc:0.89880 train_test_gap:0.02307
epoch:0 steps:293 Train_Loss:1.18462 Test_Loss:1.34422 Train_Acc:0.91406 Test_Acc:0.90930 train_test_gap:0.00476
epoch:0 steps:294 Train_Loss:2.17725 Test_Loss:1.27226 Train_Acc:0.85938 Test_Acc:0.91460 train_test_gap:-0.05522
epoch:0 steps:295 Train_Loss:1.86906 Test_Loss:1.58507 Train_Acc:0.88281 Test_Acc:0.89190 train_test_gap:-0.00909
epoch:0 steps:296 Train_Loss:2.11714 Test_Loss:1.19592 Train_Acc:0.86719 Test_Acc:0.91840 train_test_gap:-0.05121
epoch:0 steps:297 Train_Loss:0.76380 Test_Loss:1.16603 Train_Acc:0.94531 Test_Acc:0.92070 train_test_gap:0.02461
epoch:0 steps:298 Train_Loss:1.33254 Test_Loss:1.08285 Train_Acc:0.91406 Test_Acc:0.92640 train_test_gap:-0.01234
epoch:0 steps:299 Train_Loss:0.65551 Test_Loss:1.08441 Train_Acc:0.95312 Test_Acc:0.92610 train_test_gap:0.02702
epoch:0 steps:300 Train_Loss:1.46376 Test_Loss:1.03813 Train_Acc:0.90625 Test_Acc:0.92960 train_test_gap:-0.02335
epoch:0 steps:301 Train_Loss:1.26361 Test_Loss:1.07924 Train_Acc:0.91406 Test_Acc:0.92620 train_test_gap:-0.01214
epoch:0 steps:302 Train_Loss:0.97121 Test_Loss:1.02043 Train_Acc:0.92969 Test_Acc:0.92970 train_test_gap:-0.00001
epoch:0 steps:303 Train_Loss:1.19026 Test_Loss:1.01369 Train_Acc:0.92188 Test_Acc:0.92980 train_test_gap:-0.00792
epoch:0 steps:304 Train_Loss:0.94741 Test_Loss:1.13552 Train_Acc:0.92969 Test_Acc:0.92240 train_test_gap:0.00729
epoch:0 steps:305 Train_Loss:0.50398 Test_Loss:1.16150 Train_Acc:0.96875 Test_Acc:0.91960 train_test_gap:0.04915
epoch:0 steps:306 Train_Loss:0.92704 Test_Loss:1.04005 Train_Acc:0.93750 Test_Acc:0.92870 train_test_gap:0.00880
epoch:0 steps:307 Train_Loss:0.70189 Test_Loss:0.98816 Train_Acc:0.95312 Test_Acc:0.93230 train_test_gap:0.02082
epoch:0 steps:308 Train_Loss:0.75554 Test_Loss:1.02121 Train_Acc:0.95312 Test_Acc:0.93040 train_test_gap:0.02272
epoch:0 steps:309 Train_Loss:2.07732 Test_Loss:1.34873 Train_Acc:0.85938 Test_Acc:0.90810 train_test_gap:-0.04873
epoch:0 steps:310 Train_Loss:1.13330 Test_Loss:1.27921 Train_Acc:0.92969 Test_Acc:0.91280 train_test_gap:0.01689
epoch:0 steps:311 Train_Loss:1.88515 Test_Loss:1.25469 Train_Acc:0.88281 Test_Acc:0.91400 train_test_gap:-0.03119
epoch:0 steps:312 Train_Loss:1.33164 Test_Loss:1.06229 Train_Acc:0.90625 Test_Acc:0.92840 train_test_gap:-0.02215
epoch:0 steps:313 Train_Loss:1.25896 Test_Loss:1.28393 Train_Acc:0.92188 Test_Acc:0.91320 train_test_gap:0.00867
epoch:0 steps:314 Train_Loss:1.54156 Test_Loss:1.10725 Train_Acc:0.89844 Test_Acc:0.92420 train_test_gap:-0.02576
epoch:0 steps:315 Train_Loss:0.79257 Test_Loss:1.13100 Train_Acc:0.93750 Test_Acc:0.92350 train_test_gap:0.01400
epoch:0 steps:316 Train_Loss:1.25856 Test_Loss:1.09803 Train_Acc:0.92188 Test_Acc:0.92600 train_test_gap:-0.00413
epoch:0 steps:317 Train_Loss:1.00761 Test_Loss:1.06857 Train_Acc:0.93750 Test_Acc:0.92770 train_test_gap:0.00980
epoch:0 steps:318 Train_Loss:1.34204 Test_Loss:1.20250 Train_Acc:0.91406 Test_Acc:0.91750 train_test_gap:-0.00344
epoch:0 steps:319 Train_Loss:0.92371 Test_Loss:0.95618 Train_Acc:0.93750 Test_Acc:0.93420 train_test_gap:0.00330
epoch:0 steps:320 Train_Loss:0.88145 Test_Loss:1.00446 Train_Acc:0.94531 Test_Acc:0.93100 train_test_gap:0.01431
epoch:0 steps:321 Train_Loss:0.49014 Test_Loss:1.00791 Train_Acc:0.96875 Test_Acc:0.93090 train_test_gap:0.03785
epoch:0 steps:322 Train_Loss:1.25970 Test_Loss:0.94365 Train_Acc:0.92188 Test_Acc:0.93570 train_test_gap:-0.01382
epoch:0 steps:323 Train_Loss:0.89022 Test_Loss:0.95231 Train_Acc:0.93750 Test_Acc:0.93500 train_test_gap:0.00250
epoch:0 steps:324 Train_Loss:2.21522 Test_Loss:1.43998 Train_Acc:0.85938 Test_Acc:0.90230 train_test_gap:-0.04292
epoch:0 steps:325 Train_Loss:2.26661 Test_Loss:1.12157 Train_Acc:0.85938 Test_Acc:0.92360 train_test_gap:-0.06422
epoch:0 steps:326 Train_Loss:1.02225 Test_Loss:1.03857 Train_Acc:0.92969 Test_Acc:0.92940 train_test_gap:0.00029
epoch:0 steps:327 Train_Loss:0.99250 Test_Loss:1.06002 Train_Acc:0.93750 Test_Acc:0.92770 train_test_gap:0.00980
epoch:0 steps:328 Train_Loss:1.27029 Test_Loss:1.03533 Train_Acc:0.91406 Test_Acc:0.92920 train_test_gap:-0.01514
epoch:0 steps:329 Train_Loss:1.47727 Test_Loss:1.13471 Train_Acc:0.89844 Test_Acc:0.92310 train_test_gap:-0.02466
epoch:0 steps:330 Train_Loss:1.16024 Test_Loss:1.07244 Train_Acc:0.92188 Test_Acc:0.92670 train_test_gap:-0.00482
epoch:0 steps:331 Train_Loss:1.29576 Test_Loss:1.30623 Train_Acc:0.91406 Test_Acc:0.91100 train_test_gap:0.00306
epoch:0 steps:332 Train_Loss:1.38868 Test_Loss:1.16794 Train_Acc:0.91406 Test_Acc:0.92080 train_test_gap:-0.00674
epoch:0 steps:333 Train_Loss:1.25931 Test_Loss:1.05079 Train_Acc:0.92188 Test_Acc:0.92780 train_test_gap:-0.00592
epoch:0 steps:334 Train_Loss:1.21249 Test_Loss:1.36061 Train_Acc:0.92188 Test_Acc:0.90750 train_test_gap:0.01438
epoch:0 steps:335 Train_Loss:1.53853 Test_Loss:1.06636 Train_Acc:0.89062 Test_Acc:0.92660 train_test_gap:-0.03597
epoch:0 steps:336 Train_Loss:1.38937 Test_Loss:1.02976 Train_Acc:0.91406 Test_Acc:0.92890 train_test_gap:-0.01484
epoch:0 steps:337 Train_Loss:1.45777 Test_Loss:1.14207 Train_Acc:0.89844 Test_Acc:0.92140 train_test_gap:-0.02296
epoch:0 steps:338 Train_Loss:1.35581 Test_Loss:1.70953 Train_Acc:0.90625 Test_Acc:0.88280 train_test_gap:0.02345
epoch:0 steps:339 Train_Loss:1.56932 Test_Loss:1.03928 Train_Acc:0.89844 Test_Acc:0.92860 train_test_gap:-0.03016
epoch:0 steps:340 Train_Loss:0.43311 Test_Loss:1.06635 Train_Acc:0.96875 Test_Acc:0.92750 train_test_gap:0.04125
epoch:0 steps:341 Train_Loss:0.97320 Test_Loss:1.07420 Train_Acc:0.93750 Test_Acc:0.92500 train_test_gap:0.01250
epoch:0 steps:342 Train_Loss:0.75553 Test_Loss:1.16560 Train_Acc:0.95312 Test_Acc:0.91900 train_test_gap:0.03412
epoch:0 steps:343 Train_Loss:0.73426 Test_Loss:1.16065 Train_Acc:0.95312 Test_Acc:0.92050 train_test_gap:0.03263
epoch:0 steps:344 Train_Loss:0.62963 Test_Loss:1.17445 Train_Acc:0.96094 Test_Acc:0.91950 train_test_gap:0.04144
epoch:0 steps:345 Train_Loss:1.43672 Test_Loss:1.14511 Train_Acc:0.90625 Test_Acc:0.92100 train_test_gap:-0.01475
epoch:0 steps:346 Train_Loss:1.47042 Test_Loss:1.10309 Train_Acc:0.90625 Test_Acc:0.92560 train_test_gap:-0.01935
epoch:0 steps:347 Train_Loss:1.25922 Test_Loss:1.26577 Train_Acc:0.92188 Test_Acc:0.91390 train_test_gap:0.00797
epoch:0 steps:348 Train_Loss:0.87196 Test_Loss:1.07462 Train_Acc:0.92188 Test_Acc:0.92740 train_test_gap:-0.00553
epoch:0 steps:349 Train_Loss:0.75554 Test_Loss:1.08172 Train_Acc:0.95312 Test_Acc:0.92630 train_test_gap:0.02682
epoch:0 steps:350 Train_Loss:0.94117 Test_Loss:1.07148 Train_Acc:0.93750 Test_Acc:0.92560 train_test_gap:0.01190
epoch:0 steps:351 Train_Loss:1.26094 Test_Loss:1.13148 Train_Acc:0.92188 Test_Acc:0.92170 train_test_gap:0.00018
epoch:0 steps:352 Train_Loss:1.13606 Test_Loss:1.06649 Train_Acc:0.92188 Test_Acc:0.92660 train_test_gap:-0.00472
epoch:0 steps:353 Train_Loss:1.73161 Test_Loss:1.18705 Train_Acc:0.87500 Test_Acc:0.91880 train_test_gap:-0.04380
epoch:0 steps:354 Train_Loss:0.88146 Test_Loss:1.06471 Train_Acc:0.94531 Test_Acc:0.92710 train_test_gap:0.01821
epoch:0 steps:355 Train_Loss:0.88075 Test_Loss:1.07923 Train_Acc:0.94531 Test_Acc:0.92610 train_test_gap:0.01921
epoch:0 steps:356 Train_Loss:1.04292 Test_Loss:0.96652 Train_Acc:0.92969 Test_Acc:0.93370 train_test_gap:-0.00401
epoch:0 steps:357 Train_Loss:1.67995 Test_Loss:1.00699 Train_Acc:0.88281 Test_Acc:0.93110 train_test_gap:-0.04829
epoch:0 steps:358 Train_Loss:1.63685 Test_Loss:0.98109 Train_Acc:0.89844 Test_Acc:0.93210 train_test_gap:-0.03366
epoch:0 steps:359 Train_Loss:1.10836 Test_Loss:1.03330 Train_Acc:0.92188 Test_Acc:0.92890 train_test_gap:-0.00702
epoch:0 steps:360 Train_Loss:1.13330 Test_Loss:0.98027 Train_Acc:0.92969 Test_Acc:0.93220 train_test_gap:-0.00251
epoch:0 steps:361 Train_Loss:1.22785 Test_Loss:1.24601 Train_Acc:0.89844 Test_Acc:0.91460 train_test_gap:-0.01616
epoch:0 steps:362 Train_Loss:1.30582 Test_Loss:1.01487 Train_Acc:0.91406 Test_Acc:0.93060 train_test_gap:-0.01654
epoch:0 steps:363 Train_Loss:2.39253 Test_Loss:1.21925 Train_Acc:0.85156 Test_Acc:0.91590 train_test_gap:-0.06434
epoch:0 steps:364 Train_Loss:2.17579 Test_Loss:1.34246 Train_Acc:0.85938 Test_Acc:0.90790 train_test_gap:-0.04853
epoch:0 steps:365 Train_Loss:0.84014 Test_Loss:1.12483 Train_Acc:0.92969 Test_Acc:0.92360 train_test_gap:0.00609
epoch:0 steps:366 Train_Loss:0.97631 Test_Loss:1.19560 Train_Acc:0.93750 Test_Acc:0.91910 train_test_gap:0.01840
epoch:0 steps:367 Train_Loss:0.83987 Test_Loss:1.25044 Train_Acc:0.93750 Test_Acc:0.91420 train_test_gap:0.02330
epoch:0 steps:368 Train_Loss:1.00731 Test_Loss:1.03730 Train_Acc:0.93750 Test_Acc:0.92840 train_test_gap:0.00910
epoch:0 steps:369 Train_Loss:1.42671 Test_Loss:1.14045 Train_Acc:0.90625 Test_Acc:0.92190 train_test_gap:-0.01565
epoch:0 steps:370 Train_Loss:0.89472 Test_Loss:1.15529 Train_Acc:0.93750 Test_Acc:0.92100 train_test_gap:0.01650
epoch:0 steps:371 Train_Loss:1.49700 Test_Loss:1.07933 Train_Acc:0.90625 Test_Acc:0.92500 train_test_gap:-0.01875
epoch:0 steps:372 Train_Loss:1.35788 Test_Loss:1.14675 Train_Acc:0.91406 Test_Acc:0.92250 train_test_gap:-0.00844
epoch:0 steps:373 Train_Loss:1.15312 Test_Loss:1.11662 Train_Acc:0.91406 Test_Acc:0.92230 train_test_gap:-0.00824
epoch:0 steps:374 Train_Loss:0.69254 Test_Loss:1.19143 Train_Acc:0.95312 Test_Acc:0.91890 train_test_gap:0.03422
epoch:0 steps:375 Train_Loss:0.76550 Test_Loss:1.06165 Train_Acc:0.94531 Test_Acc:0.92660 train_test_gap:0.01871
epoch:0 steps:376 Train_Loss:1.25923 Test_Loss:1.14358 Train_Acc:0.92188 Test_Acc:0.92100 train_test_gap:0.00087
epoch:0 steps:377 Train_Loss:1.51120 Test_Loss:1.40397 Train_Acc:0.90625 Test_Acc:0.90550 train_test_gap:0.00075
epoch:0 steps:378 Train_Loss:0.55316 Test_Loss:1.01268 Train_Acc:0.95312 Test_Acc:0.93040 train_test_gap:0.02272
epoch:0 steps:379 Train_Loss:1.56579 Test_Loss:1.17784 Train_Acc:0.89062 Test_Acc:0.91950 train_test_gap:-0.02887
epoch:0 steps:380 Train_Loss:0.46709 Test_Loss:1.08977 Train_Acc:0.96875 Test_Acc:0.92490 train_test_gap:0.04385
epoch:0 steps:381 Train_Loss:1.24444 Test_Loss:1.14599 Train_Acc:0.92188 Test_Acc:0.92240 train_test_gap:-0.00052
epoch:0 steps:382 Train_Loss:1.12438 Test_Loss:1.12511 Train_Acc:0.92188 Test_Acc:0.92310 train_test_gap:-0.00123
epoch:0 steps:383 Train_Loss:0.37777 Test_Loss:1.12016 Train_Acc:0.97656 Test_Acc:0.92280 train_test_gap:0.05376
epoch:0 steps:384 Train_Loss:2.32186 Test_Loss:1.87083 Train_Acc:0.85156 Test_Acc:0.87420 train_test_gap:-0.02264
epoch:0 steps:385 Train_Loss:2.30597 Test_Loss:1.28650 Train_Acc:0.84375 Test_Acc:0.91280 train_test_gap:-0.06905
epoch:0 steps:386 Train_Loss:1.79826 Test_Loss:1.04203 Train_Acc:0.88281 Test_Acc:0.92820 train_test_gap:-0.04539
epoch:0 steps:387 Train_Loss:0.83334 Test_Loss:0.99424 Train_Acc:0.93750 Test_Acc:0.93190 train_test_gap:0.00560
epoch:0 steps:388 Train_Loss:1.38412 Test_Loss:1.12973 Train_Acc:0.91406 Test_Acc:0.92310 train_test_gap:-0.00904
epoch:0 steps:389 Train_Loss:1.59948 Test_Loss:1.10247 Train_Acc:0.89844 Test_Acc:0.92410 train_test_gap:-0.02566
epoch:0 steps:390 Train_Loss:1.26623 Test_Loss:1.05284 Train_Acc:0.91406 Test_Acc:0.92790 train_test_gap:-0.01384
epoch:0 steps:391 Train_Loss:1.56317 Test_Loss:1.17915 Train_Acc:0.89062 Test_Acc:0.91970 train_test_gap:-0.02907
epoch:0 steps:392 Train_Loss:0.79966 Test_Loss:1.04832 Train_Acc:0.94531 Test_Acc:0.92800 train_test_gap:0.01731
epoch:0 steps:393 Train_Loss:1.13102 Test_Loss:1.11628 Train_Acc:0.91406 Test_Acc:0.92230 train_test_gap:-0.00824
epoch:0 steps:394 Train_Loss:1.38491 Test_Loss:1.05230 Train_Acc:0.91406 Test_Acc:0.92790 train_test_gap:-0.01384
epoch:0 steps:395 Train_Loss:1.98679 Test_Loss:1.34273 Train_Acc:0.87500 Test_Acc:0.91000 train_test_gap:-0.03500
epoch:0 steps:396 Train_Loss:1.14413 Test_Loss:1.15613 Train_Acc:0.92188 Test_Acc:0.92220 train_test_gap:-0.00033
epoch:0 steps:397 Train_Loss:1.30336 Test_Loss:1.04164 Train_Acc:0.91406 Test_Acc:0.92820 train_test_gap:-0.01414
epoch:0 steps:398 Train_Loss:1.15242 Test_Loss:1.12324 Train_Acc:0.92188 Test_Acc:0.92280 train_test_gap:-0.00092
epoch:0 steps:399 Train_Loss:0.48362 Test_Loss:1.09722 Train_Acc:0.96875 Test_Acc:0.92290 train_test_gap:0.04585
epoch:0 steps:400 Train_Loss:0.66576 Test_Loss:0.93220 Train_Acc:0.94531 Test_Acc:0.93490 train_test_gap:0.01041
epoch:0 steps:401 Train_Loss:0.83125 Test_Loss:0.94781 Train_Acc:0.94531 Test_Acc:0.93530 train_test_gap:0.01001
epoch:0 steps:402 Train_Loss:1.65144 Test_Loss:1.10800 Train_Acc:0.89062 Test_Acc:0.92420 train_test_gap:-0.03358
epoch:0 steps:403 Train_Loss:0.88146 Test_Loss:0.97099 Train_Acc:0.94531 Test_Acc:0.93280 train_test_gap:0.01251
epoch:0 steps:404 Train_Loss:0.92285 Test_Loss:0.97724 Train_Acc:0.93750 Test_Acc:0.93180 train_test_gap:0.00570
epoch:0 steps:405 Train_Loss:0.88146 Test_Loss:1.00943 Train_Acc:0.94531 Test_Acc:0.93060 train_test_gap:0.01471
epoch:0 steps:406 Train_Loss:0.50355 Test_Loss:0.99001 Train_Acc:0.96875 Test_Acc:0.93110 train_test_gap:0.03765
epoch:0 steps:407 Train_Loss:0.50369 Test_Loss:1.08708 Train_Acc:0.96875 Test_Acc:0.92510 train_test_gap:0.04365
epoch:0 steps:408 Train_Loss:1.39747 Test_Loss:0.96063 Train_Acc:0.90625 Test_Acc:0.93310 train_test_gap:-0.02685
epoch:0 steps:409 Train_Loss:1.84178 Test_Loss:1.08598 Train_Acc:0.88281 Test_Acc:0.92520 train_test_gap:-0.04239
epoch:0 steps:410 Train_Loss:1.19580 Test_Loss:1.12391 Train_Acc:0.92188 Test_Acc:0.92270 train_test_gap:-0.00082
epoch:0 steps:411 Train_Loss:1.00738 Test_Loss:0.97227 Train_Acc:0.93750 Test_Acc:0.93210 train_test_gap:0.00540
epoch:0 steps:412 Train_Loss:0.50369 Test_Loss:0.97689 Train_Acc:0.96875 Test_Acc:0.93280 train_test_gap:0.03595
epoch:0 steps:413 Train_Loss:1.00712 Test_Loss:1.00463 Train_Acc:0.93750 Test_Acc:0.93150 train_test_gap:0.00600
epoch:0 steps:414 Train_Loss:1.25148 Test_Loss:1.03153 Train_Acc:0.92188 Test_Acc:0.92870 train_test_gap:-0.00682
epoch:0 steps:415 Train_Loss:2.86669 Test_Loss:1.34664 Train_Acc:0.80469 Test_Acc:0.90940 train_test_gap:-0.10471
epoch:0 steps:416 Train_Loss:1.01365 Test_Loss:1.01745 Train_Acc:0.92188 Test_Acc:0.93100 train_test_gap:-0.00913
epoch:0 steps:417 Train_Loss:0.75554 Test_Loss:0.99708 Train_Acc:0.95312 Test_Acc:0.93260 train_test_gap:0.02053
epoch:0 steps:418 Train_Loss:0.36324 Test_Loss:1.03857 Train_Acc:0.97656 Test_Acc:0.93010 train_test_gap:0.04646
epoch:0 steps:419 Train_Loss:0.64857 Test_Loss:0.97160 Train_Acc:0.95312 Test_Acc:0.93350 train_test_gap:0.01963
epoch:0 steps:420 Train_Loss:0.82359 Test_Loss:0.98567 Train_Acc:0.94531 Test_Acc:0.93200 train_test_gap:0.01331
epoch:0 steps:421 Train_Loss:1.13338 Test_Loss:0.89866 Train_Acc:0.92969 Test_Acc:0.93810 train_test_gap:-0.00841
epoch:0 steps:422 Train_Loss:0.79081 Test_Loss:0.99931 Train_Acc:0.94531 Test_Acc:0.93090 train_test_gap:0.01441
epoch:0 steps:423 Train_Loss:0.96477 Test_Loss:1.02616 Train_Acc:0.92188 Test_Acc:0.92940 train_test_gap:-0.00753
epoch:0 steps:424 Train_Loss:0.93060 Test_Loss:0.95378 Train_Acc:0.92188 Test_Acc:0.93400 train_test_gap:-0.01213
epoch:0 steps:425 Train_Loss:0.88124 Test_Loss:0.94577 Train_Acc:0.94531 Test_Acc:0.93450 train_test_gap:0.01081
epoch:0 steps:426 Train_Loss:0.72347 Test_Loss:0.97409 Train_Acc:0.93750 Test_Acc:0.93230 train_test_gap:0.00520
epoch:0 steps:427 Train_Loss:0.63109 Test_Loss:1.03073 Train_Acc:0.96094 Test_Acc:0.92980 train_test_gap:0.03114
epoch:0 steps:428 Train_Loss:0.81403 Test_Loss:0.96041 Train_Acc:0.94531 Test_Acc:0.93310 train_test_gap:0.01221
epoch:0 steps:429 Train_Loss:0.72472 Test_Loss:1.06760 Train_Acc:0.94531 Test_Acc:0.92610 train_test_gap:0.01921
epoch:0 steps:430 Train_Loss:1.81829 Test_Loss:1.02954 Train_Acc:0.88281 Test_Acc:0.92790 train_test_gap:-0.04509
epoch:0 steps:431 Train_Loss:1.03501 Test_Loss:1.15035 Train_Acc:0.92969 Test_Acc:0.92130 train_test_gap:0.00839
epoch:0 steps:432 Train_Loss:0.50369 Test_Loss:1.11432 Train_Acc:0.96875 Test_Acc:0.92360 train_test_gap:0.04515
epoch:0 steps:433 Train_Loss:1.41516 Test_Loss:1.13014 Train_Acc:0.90625 Test_Acc:0.92140 train_test_gap:-0.01515
epoch:0 steps:434 Train_Loss:1.34740 Test_Loss:0.91847 Train_Acc:0.91406 Test_Acc:0.93640 train_test_gap:-0.02234
epoch:0 steps:435 Train_Loss:1.24241 Test_Loss:1.02409 Train_Acc:0.92188 Test_Acc:0.92930 train_test_gap:-0.00743
epoch:0 steps:436 Train_Loss:0.37776 Test_Loss:0.92271 Train_Acc:0.97656 Test_Acc:0.93580 train_test_gap:0.04076
epoch:0 steps:437 Train_Loss:0.91897 Test_Loss:0.89734 Train_Acc:0.92969 Test_Acc:0.93820 train_test_gap:-0.00851
epoch:0 steps:438 Train_Loss:1.12225 Test_Loss:0.93269 Train_Acc:0.91406 Test_Acc:0.93540 train_test_gap:-0.02134
epoch:0 steps:439 Train_Loss:0.49416 Test_Loss:0.85613 Train_Acc:0.96875 Test_Acc:0.94090 train_test_gap:0.02785
epoch:0 steps:440 Train_Loss:0.62964 Test_Loss:0.88893 Train_Acc:0.96094 Test_Acc:0.93820 train_test_gap:0.02274
epoch:0 steps:441 Train_Loss:0.95380 Test_Loss:0.89211 Train_Acc:0.93750 Test_Acc:0.93850 train_test_gap:-0.00100
epoch:0 steps:442 Train_Loss:0.88841 Test_Loss:0.83500 Train_Acc:0.93750 Test_Acc:0.94170 train_test_gap:-0.00420
epoch:0 steps:443 Train_Loss:0.50119 Test_Loss:0.88813 Train_Acc:0.96875 Test_Acc:0.93870 train_test_gap:0.03005
epoch:0 steps:444 Train_Loss:0.93268 Test_Loss:0.96942 Train_Acc:0.93750 Test_Acc:0.93330 train_test_gap:0.00420
epoch:0 steps:445 Train_Loss:0.62966 Test_Loss:0.89181 Train_Acc:0.96094 Test_Acc:0.93820 train_test_gap:0.02274
epoch:0 steps:446 Train_Loss:0.37777 Test_Loss:0.92085 Train_Acc:0.97656 Test_Acc:0.93570 train_test_gap:0.04086
epoch:0 steps:447 Train_Loss:0.75189 Test_Loss:0.92848 Train_Acc:0.95312 Test_Acc:0.93550 train_test_gap:0.01763
epoch:0 steps:448 Train_Loss:0.49905 Test_Loss:0.94747 Train_Acc:0.96875 Test_Acc:0.93520 train_test_gap:0.03355
epoch:0 steps:449 Train_Loss:1.30218 Test_Loss:1.16318 Train_Acc:0.91406 Test_Acc:0.92030 train_test_gap:-0.00624
epoch:0 steps:450 Train_Loss:0.84515 Test_Loss:0.92141 Train_Acc:0.92969 Test_Acc:0.93650 train_test_gap:-0.00681
epoch:0 steps:451 Train_Loss:0.37777 Test_Loss:0.90761 Train_Acc:0.97656 Test_Acc:0.93720 train_test_gap:0.03936
epoch:0 steps:452 Train_Loss:1.13327 Test_Loss:0.91310 Train_Acc:0.92969 Test_Acc:0.93720 train_test_gap:-0.00751
epoch:0 steps:453 Train_Loss:1.44538 Test_Loss:1.08710 Train_Acc:0.89844 Test_Acc:0.92580 train_test_gap:-0.02736
epoch:0 steps:454 Train_Loss:1.00846 Test_Loss:0.90418 Train_Acc:0.93750 Test_Acc:0.93730 train_test_gap:0.00020
epoch:0 steps:455 Train_Loss:0.75558 Test_Loss:0.91758 Train_Acc:0.95312 Test_Acc:0.93650 train_test_gap:0.01663
epoch:0 steps:456 Train_Loss:0.29716 Test_Loss:0.93556 Train_Acc:0.97656 Test_Acc:0.93520 train_test_gap:0.04136
epoch:0 steps:457 Train_Loss:0.37777 Test_Loss:0.99258 Train_Acc:0.97656 Test_Acc:0.93210 train_test_gap:0.04446
epoch:0 steps:458 Train_Loss:0.56498 Test_Loss:0.93960 Train_Acc:0.96094 Test_Acc:0.93540 train_test_gap:0.02554
epoch:0 steps:459 Train_Loss:0.50369 Test_Loss:0.89870 Train_Acc:0.96875 Test_Acc:0.93740 train_test_gap:0.03135
epoch:0 steps:460 Train_Loss:0.50335 Test_Loss:0.90137 Train_Acc:0.96875 Test_Acc:0.93740 train_test_gap:0.03135
epoch:0 steps:461 Train_Loss:1.01954 Test_Loss:0.95797 Train_Acc:0.92969 Test_Acc:0.93490 train_test_gap:-0.00521
epoch:0 steps:462 Train_Loss:0.16159 Test_Loss:0.99130 Train_Acc:0.98438 Test_Acc:0.93230 train_test_gap:0.05207
epoch:0 steps:463 Train_Loss:0.12592 Test_Loss:0.95475 Train_Acc:0.99219 Test_Acc:0.93430 train_test_gap:0.05789
epoch:0 steps:464 Train_Loss:0.65007 Test_Loss:1.16769 Train_Acc:0.95312 Test_Acc:0.91920 train_test_gap:0.03392
epoch:0 steps:465 Train_Loss:1.86314 Test_Loss:1.02700 Train_Acc:0.87500 Test_Acc:0.92950 train_test_gap:-0.05450
epoch:0 steps:466 Train_Loss:0.50551 Test_Loss:0.94951 Train_Acc:0.96875 Test_Acc:0.93430 train_test_gap:0.03445
epoch:0 steps:467 Train_Loss:0.00189 Test_Loss:0.95302 Train_Acc:1.00000 Test_Acc:0.93430 train_test_gap:0.06570
epoch:0 steps:468 Train_Loss:1.76292 Test_Loss:1.01102 Train_Acc:0.89062 Test_Acc:0.93060 train_test_gap:-0.03997
epoch:0 steps:469 Train_Loss:0.12593 Test_Loss:1.04530 Train_Acc:0.99219 Test_Acc:0.92890 train_test_gap:0.06329
epoch:1 steps:470 Train_Loss:4.91182 Test_Loss:1.17441 Train_Acc:0.68750 Test_Acc:0.92040 train_test_gap:-0.23290
epoch:1 steps:471 Train_Loss:0.50389 Test_Loss:1.06886 Train_Acc:0.96875 Test_Acc:0.92610 train_test_gap:0.04265
epoch:1 steps:472 Train_Loss:1.00738 Test_Loss:0.96774 Train_Acc:0.93750 Test_Acc:0.93280 train_test_gap:0.00470
epoch:1 steps:473 Train_Loss:1.50204 Test_Loss:1.14688 Train_Acc:0.90625 Test_Acc:0.92190 train_test_gap:-0.01565
epoch:1 steps:474 Train_Loss:1.13469 Test_Loss:0.89937 Train_Acc:0.92188 Test_Acc:0.93820 train_test_gap:-0.01633
epoch:1 steps:475 Train_Loss:0.74542 Test_Loss:0.84396 Train_Acc:0.94531 Test_Acc:0.94150 train_test_gap:0.00381
epoch:1 steps:476 Train_Loss:0.82882 Test_Loss:0.86119 Train_Acc:0.93750 Test_Acc:0.94000 train_test_gap:-0.00250
epoch:1 steps:477 Train_Loss:1.27838 Test_Loss:0.87853 Train_Acc:0.91406 Test_Acc:0.93950 train_test_gap:-0.02544
epoch:1 steps:478 Train_Loss:1.00738 Test_Loss:0.81168 Train_Acc:0.93750 Test_Acc:0.94280 train_test_gap:-0.00530
epoch:1 steps:479 Train_Loss:1.57119 Test_Loss:0.93495 Train_Acc:0.88281 Test_Acc:0.93600 train_test_gap:-0.05319
epoch:1 steps:480 Train_Loss:0.99319 Test_Loss:0.97769 Train_Acc:0.93750 Test_Acc:0.93190 train_test_gap:0.00560
epoch:1 steps:481 Train_Loss:1.24683 Test_Loss:0.96928 Train_Acc:0.92188 Test_Acc:0.93460 train_test_gap:-0.01272
epoch:1 steps:482 Train_Loss:0.67633 Test_Loss:0.84979 Train_Acc:0.95312 Test_Acc:0.94130 train_test_gap:0.01182
epoch:1 steps:483 Train_Loss:0.78322 Test_Loss:0.87392 Train_Acc:0.94531 Test_Acc:0.93940 train_test_gap:0.00591
epoch:1 steps:484 Train_Loss:0.37777 Test_Loss:0.94248 Train_Acc:0.97656 Test_Acc:0.93510 train_test_gap:0.04146
epoch:1 steps:485 Train_Loss:0.32239 Test_Loss:0.84399 Train_Acc:0.97656 Test_Acc:0.94170 train_test_gap:0.03486
epoch:1 steps:486 Train_Loss:0.74827 Test_Loss:0.90175 Train_Acc:0.94531 Test_Acc:0.93810 train_test_gap:0.00721
epoch:1 steps:487 Train_Loss:0.66472 Test_Loss:0.86780 Train_Acc:0.95312 Test_Acc:0.93940 train_test_gap:0.01372
epoch:1 steps:488 Train_Loss:1.03737 Test_Loss:0.84049 Train_Acc:0.92969 Test_Acc:0.94130 train_test_gap:-0.01161
epoch:1 steps:489 Train_Loss:0.79868 Test_Loss:0.86942 Train_Acc:0.94531 Test_Acc:0.94060 train_test_gap:0.00471
epoch:1 steps:490 Train_Loss:0.26015 Test_Loss:0.85545 Train_Acc:0.98438 Test_Acc:0.94110 train_test_gap:0.04327
epoch:1 steps:491 Train_Loss:0.68420 Test_Loss:0.86221 Train_Acc:0.95312 Test_Acc:0.93970 train_test_gap:0.01343
epoch:1 steps:492 Train_Loss:1.00617 Test_Loss:0.91085 Train_Acc:0.93750 Test_Acc:0.93710 train_test_gap:0.00040
epoch:1 steps:493 Train_Loss:0.82963 Test_Loss:0.85967 Train_Acc:0.94531 Test_Acc:0.94090 train_test_gap:0.00441
epoch:1 steps:494 Train_Loss:1.47551 Test_Loss:0.85453 Train_Acc:0.90625 Test_Acc:0.94140 train_test_gap:-0.03515
epoch:1 steps:495 Train_Loss:0.47843 Test_Loss:0.84448 Train_Acc:0.96875 Test_Acc:0.94130 train_test_gap:0.02745
epoch:1 steps:496 Train_Loss:0.63574 Test_Loss:0.84768 Train_Acc:0.95312 Test_Acc:0.94220 train_test_gap:0.01092
epoch:1 steps:497 Train_Loss:0.73388 Test_Loss:0.86306 Train_Acc:0.93750 Test_Acc:0.94050 train_test_gap:-0.00300
epoch:1 steps:498 Train_Loss:0.49146 Test_Loss:0.91310 Train_Acc:0.96875 Test_Acc:0.93690 train_test_gap:0.03185
epoch:1 steps:499 Train_Loss:0.88151 Test_Loss:0.88154 Train_Acc:0.94531 Test_Acc:0.93810 train_test_gap:0.00721
epoch:1 steps:500 Train_Loss:1.12386 Test_Loss:0.84379 Train_Acc:0.92188 Test_Acc:0.94170 train_test_gap:-0.01982
epoch:1 steps:501 Train_Loss:0.50368 Test_Loss:0.80597 Train_Acc:0.96875 Test_Acc:0.94430 train_test_gap:0.02445

DarrenXf

关注

1
点赞
踩
5

收藏

觉得还不错? 一键收藏
0
评论
python 代码实现反向传播算法

实现反向传播的算法可以在python2以及python3中运行,在我的测试环境中可以运行.我并没有详细去测试每一个python版本.算法中使用的数据是mnist数据集.下面是算法的代码forward_neural_network.py#!/usr/bin/python# -*- coding: utf-8 -*-#####################################...
复制链接

扫一扫