空间站广场

论文

Notebooks

比赛

课程

Apps

镜像市场

实验室

Uni-Lab

我的主页

我的Notebooks

我的知识库

我的足迹

我的工作空间

任务

节点

镜像

文件

数据集

项目

数据库

公开

Machine Learning Tutorials

Machine Learning

sisyphus

更新于 2024-12-25

推荐镜像 :Basic Image:ubuntu22.04-py3.10

推荐机型 :c3_m4_1 * NVIDIA T4

数据集

covid(v1)

1. 测试环境可用性

代码

文本

[2]

import torch

x = torch.tensor([[1.,0.], [-1.,1.]], requires_grad=True)

z = x.pow(2).sum()

z.backward()

print(x.grad)

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

print(f'device: {device}')

tensor([[ 2.,  0.],
        [-2.,  2.]])
device: cuda

代码

文本

2. 依赖包引入

代码

文本

[3]

# 数据、矩阵运算

import math

import numpy as np

# 数据读写

import pandas as pd

import os

import csv

# 进度条

from tqdm.notebook import tqdm

# 深度学习

import torch

import torch.nn as nn

from torch.utils.data import DataLoader, Dataset, random_split

from torchviz import make_dot

from torch.utils.tensorboard import SummaryWriter

代码

文本

3. 训练准备

代码

文本

[4]

def same_seed(seed):

torch.backends.cudnn.deterministic = True # 保证每次结果一样

torch.backends.cudnn.benchmark = False

np.random.seed(seed)

torch.manual_seed(seed)

if torch.cuda.is_available():

torch.cuda.manual_seed_all(seed)

print(f"set seed = {seed}")

def train_valid_split(dataset, valid_ratio=0.2, seed=1):

valid_set_size = int(valid_ratio * len(dataset))

train_set_size = len(dataset) - valid_set_size

train_dataset, valid_dataset = random_split(dataset, [train_set_size, valid_set_size], generator=torch.Generator().manual_seed(seed))

return np.array(train_dataset), np.array(valid_dataset)

def predict(test_loader, model, device):

model.eval()

preds = []

for x in tqdm(test_loader):

x = x.to(device)

with torch.no_grad():

y_pred = model(x)

preds.append(y_pred.cpu())

preds = torch.cat(preds, dim=0).numpy()

return preds

代码

文本

[5]

class COVIDDataset(Dataset):

"""_summary_

Args:

Dataset (_type_): _description_

x: np.ndarray 特征矩阵

y: np.ndarray 标签向量，如果是 None 则为预测的数据集

"""

def __init__(self, x, y=None):

if y is None:

self.y = y

else:

self.y = torch.FloatTensor(y)

self.x = torch.FloatTensor(x)

def __getitem__(self, index):

if self.y is None:

return self.x[index]

return self.x[index], self.y[index]

def __len__(self):

return len(self.x)

代码

文本

[6]

class MyModel(nn.Module):

def __init__(self, input_dim, hidden_dims=[128, 256, 128], dropout_rate=0.2):

# hidden_dims 的设置能够先增加维度，再减少维度，帮助捕获 low/high level 的特征

super(MyModel, self).__init__()

# Batch Normalization for input

# 增加 normalization 层能提升训练稳定性，对时序数据尤为重要

self.input_bn = nn.BatchNorm1d(input_dim)

# Build dynamic layers

layers = []

prev_dim = input_dim

# Add hidden layers

for hidden_dim in hidden_dims:

layers.extend([

nn.Linear(prev_dim, hidden_dim),

nn.BatchNorm1d(hidden_dim),

nn.ReLU(),

nn.Dropout(dropout_rate) # 增加 dropout 层能减少过拟合，提升泛化能力

])

prev_dim = hidden_dim

self.hidden_layers = nn.Sequential(*layers)

# Output layer

self.output_layer = nn.Sequential(

nn.Linear(prev_dim, 64),

nn.ReLU(),

nn.Linear(64, 1)

)

# Optional: Residual connections

# 对较深的网络很重要，提升训练稳定性

self.has_residual = (input_dim == hidden_dims[-1])

if self.has_residual:

self.residual_proj = nn.Linear(input_dim, hidden_dims[-1])

# self.net = nn.Sequential(

# nn.Linear(input_dim, 32),

# nn.Sigmoid(),

# nn.Linear(32, 32),

# nn.Sigmoid(),

# nn.Linear(32, 1)

# )

def forward(self, x):

# x = self.net(x)

# x = x.squeeze(1)

# return x

# Input normalization

x = self.input_bn(x)

# Process through hidden layers

hidden = self.hidden_layers(x)

# Add residual connection if dimensions match

if self.has_residual:

hidden = hidden + self.residual_proj(x)

# Output processing

out = self.output_layer(hidden)

return out.squeeze(1)

代码

文本

[7]

def select_feat(train_data, valid_data, test_data, select_all=True):

y_train, y_valid = train_data[:, -1], valid_data[:, -1]

raw_x_train, raw_x_valid, raw_x_test = train_data[:, :-1], valid_data[:, :-1], test_data

if select_all:

feat_idx = list(range(raw_x_train.shape[1]))

else:

feat_idx = list(range(52)) + list(range(53, 54)) # 优化选择特征

return raw_x_train[:, feat_idx], raw_x_valid[:, feat_idx], raw_x_test[:, feat_idx], y_train, y_valid

代码

文本

[8]

def trainer(train_loader, valid_loader, model, config, device):

criterion = nn.MSELoss(reduction='mean')

# criterion = nn.BCELoss(reduction='mean') # 二分类问题使用交叉熵损失函数，适合判断 0-1 范围的数值

# criterion = nn.BCEWithLogitsLoss(reduction='mean') # 二分类问题结合 Sigmoid 激活，数值稳定性更好

# criterion = nn.HuberLoss(delta=1.0) # 平滑的 L1 损失函数，在误差较小时是 MSE，在误差较大时是 L1，更擅长处理异常值

# optimizer = torch.optim.SGD(model.parameters(), lr=config["learning_rate"], momentum=0.9)

optimizer = torch.optim.Adam(model.parameters(), lr=config["learning_rate"], weight_decay=config["weight_decay"], betas=(0.9, 0.999)) # 自适应学习率优化器，结合了梯度的均值和方差来调整每个参数的学习率，更擅长处理稀疏梯度和非平稳目标

# optimizer = torch.optim.AdamW(model.parameters(), lr=config["learning_rate"], weight_decay=config["weight_decay"]) # 自适应学习率优化器，结合了梯度的均值和方差来调整每个参数的学习率，更擅长处理稀疏梯度和非平稳目标

# scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, mode="min", factor=0.1, patience=10, verbose=True) # 学习率调度器，当验证损失不再下降时，降低学习率

writer = SummaryWriter(log_dir=config["log_dir"])

if not os.path.exists(config["log_dir"]):

os.makedirs(config["log_dir"])

if not os.path.exists(config["model_dir"]):

os.makedirs(config["model_dir"])

n_epochs, best_loss, step, early_stop_count = config["n_epochs"], math.inf, 0, 0

for epoch in range(n_epochs):

model.train()

loss_record = []

# 进度条, position=0 表示第一个进度条，leave=True 表示训练完成后不删除进度条

train_pbar = tqdm(train_loader, position=0, leave=True)

train_pbar.set_description(f"Epoch [{epoch+1}/{n_epochs}]")

for x, y in train_pbar:

optimizer.zero_grad() # 梯度清零

x, y = x.to(device), y.to(device)

y_pred = model(x)

loss = criterion(y_pred, y)

loss.backward() # 反向传播，计算梯度

optimizer.step() # 更新参数

step += 1

loss_record.append(loss.item())

# 训练完一个 batch 后将 loss 显示在进度条上，detach() 将 loss 从计算图中分离，避免内存占用

train_pbar.set_postfix({'train_loss': loss.detach().item()})

mean_train_loss = sum(loss_record) / len(loss_record)

writer.add_scalar("train_loss", mean_train_loss, epoch)

model.eval()

loss_record = []

for x, y in valid_loader:

x, y = x.to(device), y.to(device)

with torch.no_grad():

y_pred = model(x)

loss = criterion(y_pred, y)

loss_record.append(loss.item())

mean_valid_loss = sum(loss_record) / len(loss_record)

print(f'Epoch [{epoch+1}/{n_epochs}], train_loss: {mean_train_loss:.4f}, valid_loss: {mean_valid_loss:.4f}')

writer.add_scalar("valid_loss", mean_valid_loss, epoch)

if mean_valid_loss < best_loss:

best_loss = mean_valid_loss

torch.save(model.state_dict(), os.path.join(config["model_dir"], "best_model.pth"))

print(f"Save model at epoch {epoch+1} with loss {best_loss:.4f}")

early_stop_count = 0

else:

early_stop_count += 1

if early_stop_count > config["early_stop"]:

print(f"Early stop at epoch {epoch+1}")

return

代码

文本

[9]

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

config = {

"seed": 3014159,

"select_all": True, # 是否选择所有特征

"valid_ratio": 0.2,

"n_epochs": 3000, # 数据遍历训练次数

"batch_size": 256,

"learning_rate": 1e-5,

"weight_decay": 0.01,

"early_stop": 400, # 如果 early_stop 次 valid_loss 都没有下降就停止训练

"log_dir": "./logs/",

"save_path": "./models/model.ckpt", # 模型保存路径

"model_dir": "./model"

}

代码

文本

[10]

same_seed(config["seed"]) # 设置随机种子方便复现

pd.set_option('display.max_columns', 200) # 设置显示数据的列数

train_df, test_df = pd.read_csv("/bohr/covid-9qxe/v1/covid.train.csv"), pd.read_csv("/bohr/covid-9qxe/v1/covid.test.csv")

display(train_df.head(3)) # 显示数据的前三行

train_data, test_data = train_df.values, test_df.values

del train_df, test_df # 删除数据，释放内存

train_data, valid_data = train_valid_split(train_data, config["valid_ratio"], config["seed"])

print(f"""train data size: {train_data.shape}

valid data size: {valid_data.shape}

test data size: {test_data.shape}""")

# 选择特征

x_train, x_valid, x_test, y_train, y_valid = select_feat(train_data, valid_data, test_data, config["select_all"])

print(f"number of features: {x_train.shape[1]}")

train_dataset, valid_dataset, test_dataset = COVIDDataset(x_train, y_train), COVIDDataset(x_valid, y_valid), COVIDDataset(x_test, None)

train_loader = DataLoader(train_dataset, batch_size=config["batch_size"], shuffle=True, pin_memory=True)

valid_loader = DataLoader(valid_dataset, batch_size=config["batch_size"], shuffle=True, pin_memory=True)

test_loader = DataLoader(test_dataset, batch_size=config["batch_size"], shuffle=False, pin_memory=True)

set seed = 3014159

	id	CO	FL	NJ	cli	ili	hh_cmnty_cli	nohh_cmnty_cli	wearing_mask	travel_outside_state	work_outside_home	shop	restaurant	spent_time	large_event	public_transit	anxious	depressed	worried_finances	tested_positive	cli.1	ili.1	hh_cmnty_cli.1	nohh_cmnty_cli.1	wearing_mask.1	travel_outside_state.1	work_outside_home.1	shop.1	restaurant.1	spent_time.1	large_event.1	public_transit.1	anxious.1	depressed.1	worried_finances.1	tested_positive.1	cli.2	ili.2	hh_cmnty_cli.2	nohh_cmnty_cli.2	wearing_mask.2	travel_outside_state.2	work_outside_home.2	shop.2	restaurant.2	spent_time.2	large_event.2	public_transit.2	anxious.2	depressed.2	worried_finances.2	tested_positive.2	cli.3	ili.3	hh_cmnty_cli.3	nohh_cmnty_cli.3	wearing_mask.3	travel_outside_state.3	work_outside_home.3	shop.3	restaurant.3	spent_time.3	large_event.3	public_transit.3	anxious.3	depressed.3	worried_finances.3	tested_positive.3	cli.4	ili.4	hh_cmnty_cli.4	nohh_cmnty_cli.4	wearing_mask.4	travel_outside_state.4	work_outside_home.4	shop.4	restaurant.4	spent_time.4	large_event.4	public_transit.4	anxious.4	depressed.4	worried_finances.4	tested_positive.4
0	0	0	1	0	0.658466	0.724606	12.939284	8.556230	76.521693	7.863621	33.069066	64.129378	33.438714	40.539878	17.446499	3.993190	10.122837	7.841023	37.992365	7.374846	0.653157	0.713249	12.488933	8.219380	75.179558	8.083434	32.564654	65.212903	34.285572	40.697210	17.754941	3.997443	10.161621	8.067275	37.882278	7.219989	0.677783	0.737965	12.487637	8.135510	74.189695	8.001016	32.145331	66.258549	34.931095	40.889258	18.038068	3.970457	10.043014	7.983358	37.705024	7.077938	0.666751	0.723506	12.367718	8.006131	73.211902	7.873342	31.330236	66.753902	35.586606	40.741650	17.800711	3.984791	10.148898	8.262288	37.384963	7.452243	0.685628	0.740943	12.364307	8.151578	71.820231	7.944531	31.113209	67.394551	36.674291	40.743132	17.842221	4.093712	10.440071	8.627117	37.329512	7.456154
1	1	1	0	0	0.693287	0.675962	15.008467	11.091818	82.375469	11.082587	36.373964	60.471825	30.516978	43.162123	16.038761	4.519762	12.133287	10.155721	33.977549	9.850027	0.738029	0.720511	15.070049	10.990937	80.943299	10.596788	36.435828	61.653141	31.449290	43.751356	16.752518	4.385694	11.911724	9.798448	32.786802	10.050547	0.831586	0.827523	14.568504	10.578924	79.096391	10.418171	35.851804	62.835720	32.802637	43.753525	17.195922	4.671201	11.418314	10.600311	32.679030	10.388084	0.767643	0.763580	14.264862	10.212529	76.645984	10.614731	34.639961	63.652437	34.147714	44.682727	17.705140	4.938357	11.056725	10.889148	32.933142	8.707858	0.713255	0.719378	12.894363	8.919288	74.687112	10.523814	33.920257	64.398380	34.612238	44.035688	17.808103	4.924935	10.172662	9.954333	32.508881	8.010957
2	2	0	0	1	0.617041	0.617041	8.614719	5.238097	87.040999	15.569851	33.366362	58.819368	24.674416	36.678937	13.490946	6.639301	9.753290	7.691015	38.685771	3.897851	0.663597	0.635459	8.472831	5.154184	86.743467	15.915242	33.012910	60.157822	25.322821	36.456240	13.385945	6.700146	9.949982	7.833514	38.036480	3.083111	0.534394	0.497435	8.493353	5.125553	86.647289	15.963496	32.471093	60.910411	25.158967	35.968207	13.656070	6.748557	10.664876	7.975821	37.826924	3.132834	0.536224	0.498305	8.227657	5.298912	86.285798	15.788676	31.931642	61.233289	25.156025	36.032687	13.863302	6.879646	10.731187	7.847829	37.128714	3.444182	0.479111	0.432445	8.067909	5.333533	86.312419	16.477784	31.604604	62.101064	26.521875	36.746453	13.903667	7.313833	10.388712	7.956139	36.745588	2.906977

train data size: (2160, 118)
      valid data size: (539, 118)
      test data size: (1078, 117)
number of features: 117

代码

文本

4. 训练

代码

文本

[11]

model = MyModel(input_dim=x_train.shape[1]).to(device)

trainer(train_loader, valid_loader, model, config, device)

Epoch [1/3000], train_loss: 139.5766, valid_loss: 118.2404
Save model at epoch 1 with loss 118.2404

Epoch [2/3000], train_loss: 137.9884, valid_loss: 128.0933

Epoch [3/3000], train_loss: 138.3064, valid_loss: 144.4278

Epoch [4/3000], train_loss: 136.3052, valid_loss: 124.1926

Epoch [5/3000], train_loss: 134.6186, valid_loss: 117.3061
Save model at epoch 5 with loss 117.3061

Epoch [6/3000], train_loss: 133.7687, valid_loss: 121.2401

Epoch [7/3000], train_loss: 134.3205, valid_loss: 137.7359

Epoch [8/3000], train_loss: 134.7792, valid_loss: 136.8165

Epoch [9/3000], train_loss: 134.0196, valid_loss: 118.4918

Epoch [10/3000], train_loss: 133.6286, valid_loss: 131.2332

Epoch [11/3000], train_loss: 132.2152, valid_loss: 132.0747

Epoch [12/3000], train_loss: 131.9956, valid_loss: 127.4537

Epoch [13/3000], train_loss: 132.7281, valid_loss: 127.7854

Epoch [14/3000], train_loss: 131.9076, valid_loss: 127.2060

Epoch [15/3000], train_loss: 131.7397, valid_loss: 125.3474

Epoch [16/3000], train_loss: 131.0049, valid_loss: 125.6164

Epoch [17/3000], train_loss: 131.8282, valid_loss: 142.5277

Epoch [18/3000], train_loss: 129.5784, valid_loss: 127.1901

Epoch [19/3000], train_loss: 129.2432, valid_loss: 117.1289
Save model at epoch 19 with loss 117.1289

Epoch [20/3000], train_loss: 127.8943, valid_loss: 115.3251
Save model at epoch 20 with loss 115.3251

Epoch [21/3000], train_loss: 128.5448, valid_loss: 130.9576

Epoch [22/3000], train_loss: 128.9125, valid_loss: 128.8439

Epoch [23/3000], train_loss: 128.3417, valid_loss: 116.5680

Epoch [24/3000], train_loss: 127.3967, valid_loss: 131.3596

Epoch [25/3000], train_loss: 126.0662, valid_loss: 130.8689

Epoch [26/3000], train_loss: 126.5969, valid_loss: 132.4868

Epoch [27/3000], train_loss: 126.2205, valid_loss: 137.9347

Epoch [28/3000], train_loss: 126.9189, valid_loss: 124.8268

Epoch [29/3000], train_loss: 125.5745, valid_loss: 122.3595

Epoch [30/3000], train_loss: 124.6905, valid_loss: 121.4730

Epoch [31/3000], train_loss: 123.7385, valid_loss: 118.7843

Epoch [32/3000], train_loss: 124.0567, valid_loss: 118.9881

Epoch [33/3000], train_loss: 122.4790, valid_loss: 117.0455

Epoch [34/3000], train_loss: 122.5866, valid_loss: 129.2986

Epoch [35/3000], train_loss: 121.3201, valid_loss: 136.1593

Epoch [36/3000], train_loss: 121.9207, valid_loss: 120.3867

Epoch [37/3000], train_loss: 121.1384, valid_loss: 120.6967

Epoch [38/3000], train_loss: 119.6513, valid_loss: 141.1027

Epoch [39/3000], train_loss: 119.5895, valid_loss: 112.1731
Save model at epoch 39 with loss 112.1731

Epoch [40/3000], train_loss: 118.8114, valid_loss: 107.8710
Save model at epoch 40 with loss 107.8710

Epoch [41/3000], train_loss: 117.8312, valid_loss: 126.4329

Epoch [42/3000], train_loss: 118.4546, valid_loss: 111.6790

Epoch [43/3000], train_loss: 116.4630, valid_loss: 121.3394

Epoch [44/3000], train_loss: 118.5037, valid_loss: 128.4467

Epoch [45/3000], train_loss: 116.3164, valid_loss: 113.5311

Epoch [46/3000], train_loss: 115.3931, valid_loss: 119.8654

Epoch [47/3000], train_loss: 116.0778, valid_loss: 107.4870
Save model at epoch 47 with loss 107.4870

Epoch [48/3000], train_loss: 116.1162, valid_loss: 119.1821

Epoch [49/3000], train_loss: 114.7758, valid_loss: 110.1279

Epoch [50/3000], train_loss: 114.6469, valid_loss: 112.6412

Epoch [51/3000], train_loss: 112.8317, valid_loss: 110.8884

Epoch [52/3000], train_loss: 113.7202, valid_loss: 107.1872
Save model at epoch 52 with loss 107.1872

Epoch [53/3000], train_loss: 110.6876, valid_loss: 102.6585
Save model at epoch 53 with loss 102.6585

Epoch [54/3000], train_loss: 112.4359, valid_loss: 113.1940

Epoch [55/3000], train_loss: 110.5392, valid_loss: 99.7370
Save model at epoch 55 with loss 99.7370

Epoch [56/3000], train_loss: 111.7432, valid_loss: 105.4859

Epoch [57/3000], train_loss: 111.8331, valid_loss: 105.9865

Epoch [58/3000], train_loss: 109.4070, valid_loss: 113.4155

Epoch [59/3000], train_loss: 109.2464, valid_loss: 104.9797

Epoch [60/3000], train_loss: 108.7261, valid_loss: 107.0595

Epoch [61/3000], train_loss: 107.7607, valid_loss: 114.8707

Epoch [62/3000], train_loss: 107.3662, valid_loss: 99.7928

Epoch [63/3000], train_loss: 108.6218, valid_loss: 103.0415

Epoch [64/3000], train_loss: 107.3561, valid_loss: 110.9836

Epoch [65/3000], train_loss: 105.9315, valid_loss: 107.4130

Epoch [66/3000], train_loss: 104.6493, valid_loss: 99.9879

Epoch [67/3000], train_loss: 103.2033, valid_loss: 104.2499

Epoch [68/3000], train_loss: 105.1562, valid_loss: 96.0662
Save model at epoch 68 with loss 96.0662

Epoch [69/3000], train_loss: 104.0371, valid_loss: 111.0737

Epoch [70/3000], train_loss: 102.4527, valid_loss: 104.1862

Epoch [71/3000], train_loss: 102.4560, valid_loss: 109.0043

Epoch [72/3000], train_loss: 103.0405, valid_loss: 98.7913

Epoch [73/3000], train_loss: 100.3643, valid_loss: 99.2793

Epoch [74/3000], train_loss: 101.3186, valid_loss: 97.7527

Epoch [75/3000], train_loss: 99.7935, valid_loss: 106.6000

Epoch [76/3000], train_loss: 100.5092, valid_loss: 89.1752
Save model at epoch 76 with loss 89.1752

Epoch [77/3000], train_loss: 99.4333, valid_loss: 94.5176

Epoch [78/3000], train_loss: 97.9135, valid_loss: 98.1505

Epoch [79/3000], train_loss: 98.8644, valid_loss: 95.9610

Epoch [80/3000], train_loss: 98.7608, valid_loss: 92.7500

Epoch [81/3000], train_loss: 96.1122, valid_loss: 100.4405

Epoch [82/3000], train_loss: 97.3325, valid_loss: 106.7446

Epoch [83/3000], train_loss: 95.9317, valid_loss: 105.2933

Epoch [84/3000], train_loss: 93.9193, valid_loss: 94.4178

Epoch [85/3000], train_loss: 96.0619, valid_loss: 95.2798

Epoch [86/3000], train_loss: 93.7666, valid_loss: 96.6680

Epoch [87/3000], train_loss: 95.3819, valid_loss: 115.5819

Epoch [88/3000], train_loss: 95.7523, valid_loss: 93.1387

Epoch [89/3000], train_loss: 93.1239, valid_loss: 95.9260

Epoch [90/3000], train_loss: 91.6966, valid_loss: 94.1136

Epoch [91/3000], train_loss: 92.0919, valid_loss: 93.3891

Epoch [92/3000], train_loss: 92.1208, valid_loss: 84.7485
Save model at epoch 92 with loss 84.7485

Epoch [93/3000], train_loss: 90.8799, valid_loss: 86.7471

Epoch [94/3000], train_loss: 90.1810, valid_loss: 91.6646

Epoch [95/3000], train_loss: 90.8829, valid_loss: 88.7704

Epoch [96/3000], train_loss: 89.9924, valid_loss: 104.6692

Epoch [97/3000], train_loss: 88.3028, valid_loss: 90.2959

Epoch [98/3000], train_loss: 88.7687, valid_loss: 95.6125

Epoch [99/3000], train_loss: 88.5024, valid_loss: 82.5349
Save model at epoch 99 with loss 82.5349

Epoch [100/3000], train_loss: 86.6704, valid_loss: 95.1566

Epoch [101/3000], train_loss: 87.2254, valid_loss: 85.5508

Epoch [102/3000], train_loss: 87.0074, valid_loss: 85.0041

Epoch [103/3000], train_loss: 85.4205, valid_loss: 80.9055
Save model at epoch 103 with loss 80.9055

Epoch [104/3000], train_loss: 84.5719, valid_loss: 80.2012
Save model at epoch 104 with loss 80.2012

Epoch [105/3000], train_loss: 85.4982, valid_loss: 83.9355

Epoch [106/3000], train_loss: 84.2141, valid_loss: 94.0496

Epoch [107/3000], train_loss: 82.3537, valid_loss: 79.9973
Save model at epoch 107 with loss 79.9973

Epoch [108/3000], train_loss: 84.2201, valid_loss: 79.4920
Save model at epoch 108 with loss 79.4920

Epoch [109/3000], train_loss: 82.7280, valid_loss: 91.0889

Epoch [110/3000], train_loss: 82.3507, valid_loss: 77.6771
Save model at epoch 110 with loss 77.6771

Epoch [111/3000], train_loss: 81.4119, valid_loss: 81.8435

Epoch [112/3000], train_loss: 79.6351, valid_loss: 83.6556

Epoch [113/3000], train_loss: 79.9374, valid_loss: 81.4595

Epoch [114/3000], train_loss: 79.7891, valid_loss: 73.8002
Save model at epoch 114 with loss 73.8002

Epoch [115/3000], train_loss: 79.4246, valid_loss: 72.8330
Save model at epoch 115 with loss 72.8330

Epoch [116/3000], train_loss: 79.2646, valid_loss: 78.2507

Epoch [117/3000], train_loss: 77.5909, valid_loss: 77.6639

Epoch [118/3000], train_loss: 77.3433, valid_loss: 79.8459

Epoch [119/3000], train_loss: 75.8327, valid_loss: 75.3193

Epoch [120/3000], train_loss: 76.4769, valid_loss: 71.5050
Save model at epoch 120 with loss 71.5050

Epoch [121/3000], train_loss: 76.9375, valid_loss: 71.3438
Save model at epoch 121 with loss 71.3438

Epoch [122/3000], train_loss: 75.4560, valid_loss: 81.5171

Epoch [123/3000], train_loss: 75.5892, valid_loss: 78.8160

Epoch [124/3000], train_loss: 74.5354, valid_loss: 74.8508

Epoch [125/3000], train_loss: 74.4622, valid_loss: 77.1247

Epoch [126/3000], train_loss: 73.0414, valid_loss: 69.3132
Save model at epoch 126 with loss 69.3132

Epoch [127/3000], train_loss: 71.8749, valid_loss: 75.4420

Epoch [128/3000], train_loss: 73.1140, valid_loss: 71.9650

Epoch [129/3000], train_loss: 72.9915, valid_loss: 61.5677
Save model at epoch 129 with loss 61.5677

Epoch [130/3000], train_loss: 72.2881, valid_loss: 75.0963

Epoch [131/3000], train_loss: 70.3347, valid_loss: 77.9925

Epoch [132/3000], train_loss: 70.1101, valid_loss: 69.6794

Epoch [133/3000], train_loss: 68.5557, valid_loss: 70.0589

Epoch [134/3000], train_loss: 68.5729, valid_loss: 69.6483

Epoch [135/3000], train_loss: 67.2226, valid_loss: 63.6267

Epoch [136/3000], train_loss: 67.4305, valid_loss: 74.5330

Epoch [137/3000], train_loss: 68.6171, valid_loss: 68.2613

Epoch [138/3000], train_loss: 66.8302, valid_loss: 71.5435

Epoch [139/3000], train_loss: 67.5081, valid_loss: 69.6177

Epoch [140/3000], train_loss: 66.3250, valid_loss: 62.0767

Epoch [141/3000], train_loss: 65.3406, valid_loss: 64.8807

Epoch [142/3000], train_loss: 65.2325, valid_loss: 66.9634

Epoch [143/3000], train_loss: 65.4681, valid_loss: 70.8324

Epoch [144/3000], train_loss: 64.6714, valid_loss: 63.7118

Epoch [145/3000], train_loss: 62.8897, valid_loss: 65.4815

Epoch [146/3000], train_loss: 63.7216, valid_loss: 69.7600

Epoch [147/3000], train_loss: 63.1670, valid_loss: 61.4262
Save model at epoch 147 with loss 61.4262

Epoch [148/3000], train_loss: 61.6369, valid_loss: 65.2230

Epoch [149/3000], train_loss: 60.6019, valid_loss: 63.8506

Epoch [150/3000], train_loss: 60.5292, valid_loss: 62.4845

Epoch [151/3000], train_loss: 60.7085, valid_loss: 60.3764
Save model at epoch 151 with loss 60.3764

Epoch [152/3000], train_loss: 60.0858, valid_loss: 59.3511
Save model at epoch 152 with loss 59.3511

Epoch [153/3000], train_loss: 58.6724, valid_loss: 61.4382

Epoch [154/3000], train_loss: 59.0643, valid_loss: 62.5929

Epoch [155/3000], train_loss: 59.5613, valid_loss: 60.9093

Epoch [156/3000], train_loss: 58.6432, valid_loss: 57.5479
Save model at epoch 156 with loss 57.5479

Epoch [157/3000], train_loss: 57.7811, valid_loss: 58.6402

Epoch [158/3000], train_loss: 57.3099, valid_loss: 57.8240

Epoch [159/3000], train_loss: 57.1341, valid_loss: 65.9074

Epoch [160/3000], train_loss: 56.9328, valid_loss: 56.0764
Save model at epoch 160 with loss 56.0764

Epoch [161/3000], train_loss: 56.1301, valid_loss: 55.8906
Save model at epoch 161 with loss 55.8906

Epoch [162/3000], train_loss: 55.3065, valid_loss: 55.1928
Save model at epoch 162 with loss 55.1928

Epoch [163/3000], train_loss: 54.8882, valid_loss: 54.6191
Save model at epoch 163 with loss 54.6191

Epoch [164/3000], train_loss: 56.3228, valid_loss: 58.1769

Epoch [165/3000], train_loss: 54.5537, valid_loss: 53.4960
Save model at epoch 165 with loss 53.4960

Epoch [166/3000], train_loss: 53.3736, valid_loss: 54.0034

Epoch [167/3000], train_loss: 53.6187, valid_loss: 61.7457

Epoch [168/3000], train_loss: 53.4735, valid_loss: 52.5330
Save model at epoch 168 with loss 52.5330

Epoch [169/3000], train_loss: 53.0966, valid_loss: 58.3103

Epoch [170/3000], train_loss: 52.5925, valid_loss: 52.4767
Save model at epoch 170 with loss 52.4767

Epoch [171/3000], train_loss: 50.5912, valid_loss: 50.2451
Save model at epoch 171 with loss 50.2451

Epoch [172/3000], train_loss: 51.3662, valid_loss: 55.6508

Epoch [173/3000], train_loss: 50.7783, valid_loss: 53.9877

Epoch [174/3000], train_loss: 50.1967, valid_loss: 50.8341

Epoch [175/3000], train_loss: 50.0258, valid_loss: 54.7734

Epoch [176/3000], train_loss: 50.5454, valid_loss: 47.3481
Save model at epoch 176 with loss 47.3481

Epoch [177/3000], train_loss: 48.8458, valid_loss: 52.2881

Epoch [178/3000], train_loss: 49.3889, valid_loss: 54.7139

Epoch [179/3000], train_loss: 49.0955, valid_loss: 53.9006

Epoch [180/3000], train_loss: 48.1140, valid_loss: 46.1271
Save model at epoch 180 with loss 46.1271

Epoch [181/3000], train_loss: 48.2255, valid_loss: 50.0804

Epoch [182/3000], train_loss: 46.5921, valid_loss: 49.7164

Epoch [183/3000], train_loss: 47.5414, valid_loss: 49.1344

Epoch [184/3000], train_loss: 46.0906, valid_loss: 46.6167

Epoch [185/3000], train_loss: 45.6031, valid_loss: 50.1561

Epoch [186/3000], train_loss: 46.1312, valid_loss: 53.4174

Epoch [187/3000], train_loss: 47.2371, valid_loss: 49.2625

Epoch [188/3000], train_loss: 45.3723, valid_loss: 50.7213

Epoch [189/3000], train_loss: 44.5751, valid_loss: 48.1356

Epoch [190/3000], train_loss: 44.0887, valid_loss: 46.4954

Epoch [191/3000], train_loss: 44.0462, valid_loss: 46.2922

Epoch [192/3000], train_loss: 43.9364, valid_loss: 47.4322

Epoch [193/3000], train_loss: 43.5470, valid_loss: 48.0617

Epoch [194/3000], train_loss: 42.7695, valid_loss: 48.1679

Epoch [195/3000], train_loss: 42.6281, valid_loss: 43.5832
Save model at epoch 195 with loss 43.5832

Epoch [196/3000], train_loss: 42.4978, valid_loss: 44.6383

Epoch [197/3000], train_loss: 42.0846, valid_loss: 45.5380

Epoch [198/3000], train_loss: 40.9583, valid_loss: 44.3331

Epoch [199/3000], train_loss: 40.8569, valid_loss: 44.7084

Epoch [200/3000], train_loss: 40.4596, valid_loss: 42.0210
Save model at epoch 200 with loss 42.0210

Epoch [201/3000], train_loss: 40.8541, valid_loss: 40.6660
Save model at epoch 201 with loss 40.6660

Epoch [202/3000], train_loss: 39.9424, valid_loss: 44.3638

Epoch [203/3000], train_loss: 40.4781, valid_loss: 45.8559

Epoch [204/3000], train_loss: 40.3428, valid_loss: 42.6827

Epoch [205/3000], train_loss: 39.1516, valid_loss: 42.1195

Epoch [206/3000], train_loss: 39.8070, valid_loss: 42.6838

Epoch [207/3000], train_loss: 40.6520, valid_loss: 42.6091

Epoch [208/3000], train_loss: 38.6284, valid_loss: 42.6458

Epoch [209/3000], train_loss: 37.6779, valid_loss: 39.7389
Save model at epoch 209 with loss 39.7389

Epoch [210/3000], train_loss: 38.8176, valid_loss: 40.4916

Epoch [211/3000], train_loss: 37.7041, valid_loss: 39.7395

Epoch [212/3000], train_loss: 37.1442, valid_loss: 42.2820

Epoch [213/3000], train_loss: 37.1187, valid_loss: 39.1718
Save model at epoch 213 with loss 39.1718

Epoch [214/3000], train_loss: 36.9795, valid_loss: 37.8334
Save model at epoch 214 with loss 37.8334

Epoch [215/3000], train_loss: 36.5140, valid_loss: 37.0789
Save model at epoch 215 with loss 37.0789

Epoch [216/3000], train_loss: 36.6133, valid_loss: 38.0154

Epoch [217/3000], train_loss: 37.0377, valid_loss: 40.3011

Epoch [218/3000], train_loss: 35.8128, valid_loss: 39.7956

Epoch [219/3000], train_loss: 35.1184, valid_loss: 38.4814

Epoch [220/3000], train_loss: 35.5337, valid_loss: 37.3600

Epoch [221/3000], train_loss: 34.6493, valid_loss: 36.9677
Save model at epoch 221 with loss 36.9677

Epoch [222/3000], train_loss: 35.7183, valid_loss: 34.7610
Save model at epoch 222 with loss 34.7610

Epoch [223/3000], train_loss: 35.5875, valid_loss: 37.6396

Epoch [224/3000], train_loss: 34.4198, valid_loss: 37.4016

Epoch [225/3000], train_loss: 33.3776, valid_loss: 38.3435

Epoch [226/3000], train_loss: 33.3551, valid_loss: 38.6618

Epoch [227/3000], train_loss: 33.0973, valid_loss: 38.2442

Epoch [228/3000], train_loss: 33.1954, valid_loss: 37.9092

Epoch [229/3000], train_loss: 33.2633, valid_loss: 34.4034
Save model at epoch 229 with loss 34.4034

Epoch [230/3000], train_loss: 32.4083, valid_loss: 34.4970

Epoch [231/3000], train_loss: 32.3498, valid_loss: 34.7069

Epoch [232/3000], train_loss: 32.4885, valid_loss: 34.6788

Epoch [233/3000], train_loss: 31.9173, valid_loss: 36.9387

Epoch [234/3000], train_loss: 32.2846, valid_loss: 38.7241

Epoch [235/3000], train_loss: 31.7357, valid_loss: 35.5725

Epoch [236/3000], train_loss: 31.4687, valid_loss: 34.5966

Epoch [237/3000], train_loss: 31.2430, valid_loss: 34.7193

Epoch [238/3000], train_loss: 30.8314, valid_loss: 34.5142

Epoch [239/3000], train_loss: 31.1997, valid_loss: 34.9990

Epoch [240/3000], train_loss: 29.9407, valid_loss: 34.2096
Save model at epoch 240 with loss 34.2096

Epoch [241/3000], train_loss: 30.6588, valid_loss: 32.2398
Save model at epoch 241 with loss 32.2398

Epoch [242/3000], train_loss: 29.6424, valid_loss: 30.7998
Save model at epoch 242 with loss 30.7998

Epoch [243/3000], train_loss: 29.4372, valid_loss: 32.3020

Epoch [244/3000], train_loss: 29.8568, valid_loss: 33.4458

Epoch [245/3000], train_loss: 29.2109, valid_loss: 29.7223
Save model at epoch 245 with loss 29.7223

Epoch [246/3000], train_loss: 28.7886, valid_loss: 29.6441
Save model at epoch 246 with loss 29.6441

Epoch [247/3000], train_loss: 29.5827, valid_loss: 34.7528

Epoch [248/3000], train_loss: 28.9562, valid_loss: 33.2441

Epoch [249/3000], train_loss: 28.2854, valid_loss: 33.1572

Epoch [250/3000], train_loss: 28.8587, valid_loss: 30.6991

Epoch [251/3000], train_loss: 28.7662, valid_loss: 35.5481

Epoch [252/3000], train_loss: 28.0499, valid_loss: 29.2952
Save model at epoch 252 with loss 29.2952

Epoch [253/3000], train_loss: 28.5167, valid_loss: 31.4255

Epoch [254/3000], train_loss: 27.4657, valid_loss: 30.5906

Epoch [255/3000], train_loss: 27.1919, valid_loss: 27.6452
Save model at epoch 255 with loss 27.6452

Epoch [256/3000], train_loss: 27.1269, valid_loss: 31.2280

Epoch [257/3000], train_loss: 27.2889, valid_loss: 27.7559

Epoch [258/3000], train_loss: 26.7266, valid_loss: 28.1817

Epoch [259/3000], train_loss: 27.1287, valid_loss: 31.7911

Epoch [260/3000], train_loss: 26.6509, valid_loss: 28.3721

Epoch [261/3000], train_loss: 26.4960, valid_loss: 28.8733

Epoch [262/3000], train_loss: 25.9518, valid_loss: 28.8434

Epoch [263/3000], train_loss: 25.8977, valid_loss: 30.7447

Epoch [264/3000], train_loss: 25.7661, valid_loss: 25.0389
Save model at epoch 264 with loss 25.0389

Epoch [265/3000], train_loss: 25.7115, valid_loss: 27.8364

Epoch [266/3000], train_loss: 25.3538, valid_loss: 26.2311

Epoch [267/3000], train_loss: 25.4233, valid_loss: 26.0693

Epoch [268/3000], train_loss: 24.6712, valid_loss: 27.7241

Epoch [269/3000], train_loss: 24.6766, valid_loss: 27.6613

Epoch [270/3000], train_loss: 24.7760, valid_loss: 25.1794

Epoch [271/3000], train_loss: 24.1621, valid_loss: 27.7474

Epoch [272/3000], train_loss: 24.3259, valid_loss: 26.2758

Epoch [273/3000], train_loss: 24.0324, valid_loss: 24.8452
Save model at epoch 273 with loss 24.8452

Epoch [274/3000], train_loss: 23.9658, valid_loss: 25.8899

Epoch [275/3000], train_loss: 23.7692, valid_loss: 25.9323

Epoch [276/3000], train_loss: 23.9613, valid_loss: 24.3937
Save model at epoch 276 with loss 24.3937

Epoch [277/3000], train_loss: 23.9490, valid_loss: 28.5803

Epoch [278/3000], train_loss: 23.4111, valid_loss: 25.0562

Epoch [279/3000], train_loss: 22.8219, valid_loss: 24.8273

Epoch [280/3000], train_loss: 23.0853, valid_loss: 23.0868
Save model at epoch 280 with loss 23.0868

Epoch [281/3000], train_loss: 22.6869, valid_loss: 21.0552
Save model at epoch 281 with loss 21.0552

Epoch [282/3000], train_loss: 22.1655, valid_loss: 24.2188

Epoch [283/3000], train_loss: 23.3839, valid_loss: 24.4388

Epoch [284/3000], train_loss: 22.8050, valid_loss: 23.9438

Epoch [285/3000], train_loss: 22.0138, valid_loss: 22.8269

Epoch [286/3000], train_loss: 22.4008, valid_loss: 24.1586

Epoch [287/3000], train_loss: 22.1891, valid_loss: 24.2401

Epoch [288/3000], train_loss: 22.5873, valid_loss: 24.4613

Epoch [289/3000], train_loss: 21.7069, valid_loss: 23.2733

Epoch [290/3000], train_loss: 21.6300, valid_loss: 24.7138

Epoch [291/3000], train_loss: 21.1014, valid_loss: 24.2064

Epoch [292/3000], train_loss: 21.4145, valid_loss: 22.6882

Epoch [293/3000], train_loss: 20.9135, valid_loss: 23.2045

Epoch [294/3000], train_loss: 20.8766, valid_loss: 21.0190
Save model at epoch 294 with loss 21.0190

Epoch [295/3000], train_loss: 20.9150, valid_loss: 24.0186

Epoch [296/3000], train_loss: 20.5961, valid_loss: 22.0745

Epoch [297/3000], train_loss: 20.1103, valid_loss: 22.7206

Epoch [298/3000], train_loss: 20.4498, valid_loss: 22.9600

Epoch [299/3000], train_loss: 20.3236, valid_loss: 20.7816
Save model at epoch 299 with loss 20.7816

Epoch [300/3000], train_loss: 20.2211, valid_loss: 22.6592

Epoch [301/3000], train_loss: 20.3298, valid_loss: 20.9793

Epoch [302/3000], train_loss: 19.6146, valid_loss: 21.6926

Epoch [303/3000], train_loss: 19.3073, valid_loss: 19.8644
Save model at epoch 303 with loss 19.8644

Epoch [304/3000], train_loss: 19.5151, valid_loss: 20.4471

Epoch [305/3000], train_loss: 18.9155, valid_loss: 20.4127

Epoch [306/3000], train_loss: 19.1480, valid_loss: 19.6332
Save model at epoch 306 with loss 19.6332

Epoch [307/3000], train_loss: 19.1742, valid_loss: 21.0139

Epoch [308/3000], train_loss: 18.5658, valid_loss: 19.8762

Epoch [309/3000], train_loss: 19.1503, valid_loss: 21.7449

Epoch [310/3000], train_loss: 18.8438, valid_loss: 19.8752

Epoch [311/3000], train_loss: 19.0271, valid_loss: 20.5729

Epoch [312/3000], train_loss: 18.3260, valid_loss: 20.3525

Epoch [313/3000], train_loss: 18.4612, valid_loss: 22.0948

Epoch [314/3000], train_loss: 18.3451, valid_loss: 20.0307

Epoch [315/3000], train_loss: 17.9039, valid_loss: 18.5415
Save model at epoch 315 with loss 18.5415

Epoch [316/3000], train_loss: 17.5965, valid_loss: 19.1511

Epoch [317/3000], train_loss: 17.7326, valid_loss: 17.9966
Save model at epoch 317 with loss 17.9966

Epoch [318/3000], train_loss: 17.6891, valid_loss: 16.9468
Save model at epoch 318 with loss 16.9468

Epoch [319/3000], train_loss: 17.2640, valid_loss: 16.9486

Epoch [320/3000], train_loss: 17.1828, valid_loss: 19.5164

Epoch [321/3000], train_loss: 17.1324, valid_loss: 16.4671
Save model at epoch 321 with loss 16.4671

Epoch [322/3000], train_loss: 16.8391, valid_loss: 16.5769

Epoch [323/3000], train_loss: 17.2174, valid_loss: 17.6120

Epoch [324/3000], train_loss: 17.3253, valid_loss: 19.0542

Epoch [325/3000], train_loss: 17.0767, valid_loss: 19.5791

Epoch [326/3000], train_loss: 16.8368, valid_loss: 18.9674

Epoch [327/3000], train_loss: 16.4783, valid_loss: 16.9065

Epoch [328/3000], train_loss: 16.7619, valid_loss: 16.8999

Epoch [329/3000], train_loss: 16.1970, valid_loss: 16.4966

Epoch [330/3000], train_loss: 15.9637, valid_loss: 16.7539

Epoch [331/3000], train_loss: 15.9168, valid_loss: 15.7167
Save model at epoch 331 with loss 15.7167

Epoch [332/3000], train_loss: 16.2752, valid_loss: 16.9736

Epoch [333/3000], train_loss: 15.8980, valid_loss: 19.8271

Epoch [334/3000], train_loss: 15.5612, valid_loss: 16.6953

Epoch [335/3000], train_loss: 15.8085, valid_loss: 15.1529
Save model at epoch 335 with loss 15.1529

Epoch [336/3000], train_loss: 15.0833, valid_loss: 16.3495

Epoch [337/3000], train_loss: 15.3565, valid_loss: 17.0899

Epoch [338/3000], train_loss: 14.9281, valid_loss: 15.1916

Epoch [339/3000], train_loss: 14.9495, valid_loss: 17.3998

Epoch [340/3000], train_loss: 14.9180, valid_loss: 15.7117

Epoch [341/3000], train_loss: 15.8245, valid_loss: 15.1888

Epoch [342/3000], train_loss: 14.6130, valid_loss: 14.9607
Save model at epoch 342 with loss 14.9607

Epoch [343/3000], train_loss: 14.4636, valid_loss: 14.4393
Save model at epoch 343 with loss 14.4393

Epoch [344/3000], train_loss: 14.5752, valid_loss: 14.7321

Epoch [345/3000], train_loss: 14.5316, valid_loss: 16.8238

Epoch [346/3000], train_loss: 14.2919, valid_loss: 15.9079

Epoch [347/3000], train_loss: 14.0805, valid_loss: 14.3610
Save model at epoch 347 with loss 14.3610

Epoch [348/3000], train_loss: 14.6030, valid_loss: 15.0350

Epoch [349/3000], train_loss: 14.6262, valid_loss: 14.9785

Epoch [350/3000], train_loss: 14.0475, valid_loss: 13.5639
Save model at epoch 350 with loss 13.5639

Epoch [351/3000], train_loss: 13.8144, valid_loss: 13.8175

Epoch [352/3000], train_loss: 13.7548, valid_loss: 14.4437

Epoch [353/3000], train_loss: 14.1247, valid_loss: 13.6140

Epoch [354/3000], train_loss: 13.0813, valid_loss: 14.3676

Epoch [355/3000], train_loss: 13.5216, valid_loss: 14.4813

Epoch [356/3000], train_loss: 13.5596, valid_loss: 14.3970

Epoch [357/3000], train_loss: 13.3861, valid_loss: 14.4291

Epoch [358/3000], train_loss: 13.0462, valid_loss: 13.4996
Save model at epoch 358 with loss 13.4996

Epoch [359/3000], train_loss: 13.4821, valid_loss: 13.7332

Epoch [360/3000], train_loss: 13.0010, valid_loss: 12.8137
Save model at epoch 360 with loss 12.8137

Epoch [361/3000], train_loss: 13.2194, valid_loss: 13.6287

Epoch [362/3000], train_loss: 13.0679, valid_loss: 14.1793

Epoch [363/3000], train_loss: 13.0882, valid_loss: 13.5921

Epoch [364/3000], train_loss: 12.3253, valid_loss: 12.2139
Save model at epoch 364 with loss 12.2139

Epoch [365/3000], train_loss: 12.5287, valid_loss: 12.8511

Epoch [366/3000], train_loss: 12.3937, valid_loss: 13.7131

Epoch [367/3000], train_loss: 12.7380, valid_loss: 12.3571

Epoch [368/3000], train_loss: 12.1713, valid_loss: 14.3714

Epoch [369/3000], train_loss: 12.0989, valid_loss: 12.2772

Epoch [370/3000], train_loss: 11.8697, valid_loss: 12.5334

Epoch [371/3000], train_loss: 11.7762, valid_loss: 12.3503

Epoch [372/3000], train_loss: 11.7588, valid_loss: 12.0269
Save model at epoch 372 with loss 12.0269

Epoch [373/3000], train_loss: 11.5642, valid_loss: 12.6937

Epoch [374/3000], train_loss: 11.4716, valid_loss: 11.7217
Save model at epoch 374 with loss 11.7217

Epoch [375/3000], train_loss: 11.8385, valid_loss: 12.3224

Epoch [376/3000], train_loss: 11.5529, valid_loss: 11.4269
Save model at epoch 376 with loss 11.4269

Epoch [377/3000], train_loss: 11.6204, valid_loss: 12.7029

Epoch [378/3000], train_loss: 11.6361, valid_loss: 11.8900

Epoch [379/3000], train_loss: 11.6582, valid_loss: 12.2960

Epoch [380/3000], train_loss: 11.5564, valid_loss: 11.9023

Epoch [381/3000], train_loss: 10.7989, valid_loss: 10.0794
Save model at epoch 381 with loss 10.0794

Epoch [382/3000], train_loss: 11.3682, valid_loss: 10.3191

Epoch [383/3000], train_loss: 11.0321, valid_loss: 12.2517

Epoch [384/3000], train_loss: 10.8092, valid_loss: 13.1315

Epoch [385/3000], train_loss: 10.5225, valid_loss: 10.9336

Epoch [386/3000], train_loss: 11.6088, valid_loss: 11.8472

Epoch [387/3000], train_loss: 10.6487, valid_loss: 10.4222

Epoch [388/3000], train_loss: 10.4675, valid_loss: 11.6106

Epoch [389/3000], train_loss: 10.4348, valid_loss: 10.3514

Epoch [390/3000], train_loss: 10.2803, valid_loss: 10.5355

Epoch [391/3000], train_loss: 10.1977, valid_loss: 9.8180
Save model at epoch 391 with loss 9.8180

Epoch [392/3000], train_loss: 10.2696, valid_loss: 11.3423

Epoch [393/3000], train_loss: 9.9150, valid_loss: 10.9459

Epoch [394/3000], train_loss: 9.8338, valid_loss: 8.9186
Save model at epoch 394 with loss 8.9186

Epoch [395/3000], train_loss: 10.6216, valid_loss: 11.2623

Epoch [396/3000], train_loss: 10.1057, valid_loss: 10.8265

Epoch [397/3000], train_loss: 9.6757, valid_loss: 11.0251

Epoch [398/3000], train_loss: 10.5245, valid_loss: 10.7208

Epoch [399/3000], train_loss: 9.6406, valid_loss: 9.9229

Epoch [400/3000], train_loss: 9.6978, valid_loss: 10.1002

Epoch [401/3000], train_loss: 9.8200, valid_loss: 8.9470

Epoch [402/3000], train_loss: 8.9795, valid_loss: 9.6173

Epoch [403/3000], train_loss: 9.2687, valid_loss: 9.5587

Epoch [404/3000], train_loss: 9.0120, valid_loss: 9.9631

Epoch [405/3000], train_loss: 9.0390, valid_loss: 10.0724

Epoch [406/3000], train_loss: 9.3680, valid_loss: 9.8178

Epoch [407/3000], train_loss: 8.8318, valid_loss: 9.4273

Epoch [408/3000], train_loss: 8.5898, valid_loss: 10.5179

Epoch [409/3000], train_loss: 8.9025, valid_loss: 9.4861

Epoch [410/3000], train_loss: 9.3094, valid_loss: 9.2207

Epoch [411/3000], train_loss: 9.0657, valid_loss: 10.2990

Epoch [412/3000], train_loss: 9.2404, valid_loss: 9.3301

Epoch [413/3000], train_loss: 8.6734, valid_loss: 8.5486
Save model at epoch 413 with loss 8.5486

Epoch [414/3000], train_loss: 8.9605, valid_loss: 8.1316
Save model at epoch 414 with loss 8.1316

Epoch [415/3000], train_loss: 8.9426, valid_loss: 8.1333

Epoch [416/3000], train_loss: 8.2513, valid_loss: 7.7814
Save model at epoch 416 with loss 7.7814

Epoch [417/3000], train_loss: 8.8517, valid_loss: 8.5663

Epoch [418/3000], train_loss: 8.3307, valid_loss: 8.2217

Epoch [419/3000], train_loss: 8.0059, valid_loss: 7.8922

Epoch [420/3000], train_loss: 8.8909, valid_loss: 7.7573
Save model at epoch 420 with loss 7.7573

Epoch [421/3000], train_loss: 8.5841, valid_loss: 8.2409

Epoch [422/3000], train_loss: 7.6996, valid_loss: 8.3396

Epoch [423/3000], train_loss: 8.1792, valid_loss: 7.6517
Save model at epoch 423 with loss 7.6517

Epoch [424/3000], train_loss: 8.4201, valid_loss: 8.2364

Epoch [425/3000], train_loss: 7.7032, valid_loss: 7.4905
Save model at epoch 425 with loss 7.4905

Epoch [426/3000], train_loss: 7.9748, valid_loss: 7.3524
Save model at epoch 426 with loss 7.3524

Epoch [427/3000], train_loss: 7.8658, valid_loss: 7.6072

Epoch [428/3000], train_loss: 7.7183, valid_loss: 7.3604

Epoch [429/3000], train_loss: 7.6028, valid_loss: 6.8778
Save model at epoch 429 with loss 6.8778

Epoch [430/3000], train_loss: 7.6117, valid_loss: 7.1417

Epoch [431/3000], train_loss: 7.4273, valid_loss: 7.9063

Epoch [432/3000], train_loss: 7.4531, valid_loss: 6.8251
Save model at epoch 432 with loss 6.8251

Epoch [433/3000], train_loss: 7.5925, valid_loss: 8.1135

Epoch [434/3000], train_loss: 7.3048, valid_loss: 7.0983

Epoch [435/3000], train_loss: 7.2222, valid_loss: 6.9534

Epoch [436/3000], train_loss: 7.2127, valid_loss: 6.8135
Save model at epoch 436 with loss 6.8135

Epoch [437/3000], train_loss: 7.2121, valid_loss: 7.5065

Epoch [438/3000], train_loss: 7.1896, valid_loss: 7.0161

Epoch [439/3000], train_loss: 7.2970, valid_loss: 7.4038

Epoch [440/3000], train_loss: 7.2746, valid_loss: 7.0870

Epoch [441/3000], train_loss: 6.9636, valid_loss: 6.7659
Save model at epoch 441 with loss 6.7659

Epoch [442/3000], train_loss: 7.0330, valid_loss: 6.6155
Save model at epoch 442 with loss 6.6155

Epoch [443/3000], train_loss: 7.2872, valid_loss: 6.2196
Save model at epoch 443 with loss 6.2196

Epoch [444/3000], train_loss: 6.6180, valid_loss: 6.6946

Epoch [445/3000], train_loss: 6.8308, valid_loss: 6.8692

Epoch [446/3000], train_loss: 6.3433, valid_loss: 6.8782

Epoch [447/3000], train_loss: 6.9233, valid_loss: 5.9583
Save model at epoch 447 with loss 5.9583

Epoch [448/3000], train_loss: 6.5358, valid_loss: 7.1327

Epoch [449/3000], train_loss: 6.7866, valid_loss: 6.6068

Epoch [450/3000], train_loss: 6.5170, valid_loss: 6.3064

Epoch [451/3000], train_loss: 6.3133, valid_loss: 6.1815

Epoch [452/3000], train_loss: 6.4495, valid_loss: 5.8384
Save model at epoch 452 with loss 5.8384

Epoch [453/3000], train_loss: 6.3243, valid_loss: 6.1869

Epoch [454/3000], train_loss: 6.3914, valid_loss: 5.7210
Save model at epoch 454 with loss 5.7210

Epoch [455/3000], train_loss: 6.4339, valid_loss: 6.2698

Epoch [456/3000], train_loss: 6.2640, valid_loss: 6.2369

Epoch [457/3000], train_loss: 6.3955, valid_loss: 5.4040
Save model at epoch 457 with loss 5.4040

Epoch [458/3000], train_loss: 5.9724, valid_loss: 6.1047

Epoch [459/3000], train_loss: 6.0954, valid_loss: 5.9481

Epoch [460/3000], train_loss: 6.0062, valid_loss: 5.9452

Epoch [461/3000], train_loss: 6.0141, valid_loss: 5.9864

Epoch [462/3000], train_loss: 6.3069, valid_loss: 5.7081

Epoch [463/3000], train_loss: 5.9967, valid_loss: 5.3474
Save model at epoch 463 with loss 5.3474

Epoch [464/3000], train_loss: 5.7899, valid_loss: 5.5801

Epoch [465/3000], train_loss: 5.8072, valid_loss: 5.2555
Save model at epoch 465 with loss 5.2555

Epoch [466/3000], train_loss: 6.1245, valid_loss: 4.8209
Save model at epoch 466 with loss 4.8209

Epoch [467/3000], train_loss: 5.5779, valid_loss: 5.4585

Epoch [468/3000], train_loss: 6.1596, valid_loss: 4.8669

Epoch [469/3000], train_loss: 5.9678, valid_loss: 5.2497

Epoch [470/3000], train_loss: 6.1651, valid_loss: 4.5582
Save model at epoch 470 with loss 4.5582

Epoch [471/3000], train_loss: 5.4973, valid_loss: 5.2719

Epoch [472/3000], train_loss: 5.7136, valid_loss: 4.5579
Save model at epoch 472 with loss 4.5579

Epoch [473/3000], train_loss: 5.5482, valid_loss: 5.5438

Epoch [474/3000], train_loss: 5.6943, valid_loss: 4.8685

Epoch [475/3000], train_loss: 5.5733, valid_loss: 5.2196

Epoch [476/3000], train_loss: 5.2600, valid_loss: 4.6695

Epoch [477/3000], train_loss: 5.3936, valid_loss: 5.7210

Epoch [478/3000], train_loss: 5.3713, valid_loss: 4.9041

Epoch [479/3000], train_loss: 5.2774, valid_loss: 4.8763

Epoch [480/3000], train_loss: 5.2520, valid_loss: 4.8370

Epoch [481/3000], train_loss: 5.3514, valid_loss: 4.5817

Epoch [482/3000], train_loss: 5.2386, valid_loss: 5.4290

Epoch [483/3000], train_loss: 5.5448, valid_loss: 4.4739
Save model at epoch 483 with loss 4.4739

Epoch [484/3000], train_loss: 5.2942, valid_loss: 4.3961
Save model at epoch 484 with loss 4.3961

Epoch [485/3000], train_loss: 5.6434, valid_loss: 4.1755
Save model at epoch 485 with loss 4.1755

Epoch [486/3000], train_loss: 5.1696, valid_loss: 4.5617

Epoch [487/3000], train_loss: 5.6601, valid_loss: 4.9195

Epoch [488/3000], train_loss: 5.0272, valid_loss: 4.2414

Epoch [489/3000], train_loss: 5.0095, valid_loss: 4.5565

Epoch [490/3000], train_loss: 5.0918, valid_loss: 5.0330

Epoch [491/3000], train_loss: 5.0770, valid_loss: 4.0364
Save model at epoch 491 with loss 4.0364

Epoch [492/3000], train_loss: 4.6294, valid_loss: 4.4911

Epoch [493/3000], train_loss: 4.6592, valid_loss: 5.3571

Epoch [494/3000], train_loss: 4.8015, valid_loss: 4.3220

Epoch [495/3000], train_loss: 5.2263, valid_loss: 4.0054
Save model at epoch 495 with loss 4.0054

Epoch [496/3000], train_loss: 5.0947, valid_loss: 4.6748

Epoch [497/3000], train_loss: 4.7107, valid_loss: 3.7469
Save model at epoch 497 with loss 3.7469

Epoch [498/3000], train_loss: 4.9133, valid_loss: 4.2347

Epoch [499/3000], train_loss: 4.7873, valid_loss: 3.7835

Epoch [500/3000], train_loss: 4.9576, valid_loss: 4.7676

Epoch [501/3000], train_loss: 4.7156, valid_loss: 4.2729

Epoch [502/3000], train_loss: 4.7307, valid_loss: 3.9118

Epoch [503/3000], train_loss: 4.6509, valid_loss: 4.4292

Epoch [504/3000], train_loss: 4.7621, valid_loss: 3.8531

Epoch [505/3000], train_loss: 5.3312, valid_loss: 4.1102

Epoch [506/3000], train_loss: 4.4486, valid_loss: 3.6613
Save model at epoch 506 with loss 3.6613

Epoch [507/3000], train_loss: 4.4440, valid_loss: 4.0548

Epoch [508/3000], train_loss: 4.3715, valid_loss: 3.8659

Epoch [509/3000], train_loss: 4.7258, valid_loss: 3.7283

Epoch [510/3000], train_loss: 4.1733, valid_loss: 3.4206
Save model at epoch 510 with loss 3.4206

Epoch [511/3000], train_loss: 4.2755, valid_loss: 3.7753

Epoch [512/3000], train_loss: 4.5454, valid_loss: 3.9645

Epoch [513/3000], train_loss: 4.3723, valid_loss: 3.5749

Epoch [514/3000], train_loss: 4.6955, valid_loss: 3.4584

Epoch [515/3000], train_loss: 4.3437, valid_loss: 3.9729

Epoch [516/3000], train_loss: 4.4107, valid_loss: 3.7339

Epoch [517/3000], train_loss: 4.2524, valid_loss: 3.7213

Epoch [518/3000], train_loss: 4.7643, valid_loss: 3.3312
Save model at epoch 518 with loss 3.3312

Epoch [519/3000], train_loss: 4.2070, valid_loss: 3.5706

Epoch [520/3000], train_loss: 4.1614, valid_loss: 3.0374
Save model at epoch 520 with loss 3.0374

Epoch [521/3000], train_loss: 4.1150, valid_loss: 3.2976

Epoch [522/3000], train_loss: 4.0469, valid_loss: 3.8626

Epoch [523/3000], train_loss: 4.1446, valid_loss: 3.1678

Epoch [524/3000], train_loss: 4.1847, valid_loss: 2.9761
Save model at epoch 524 with loss 2.9761

Epoch [525/3000], train_loss: 4.2788, valid_loss: 4.3632

Epoch [526/3000], train_loss: 4.2572, valid_loss: 3.7291

Epoch [527/3000], train_loss: 4.3660, valid_loss: 3.4724

Epoch [528/3000], train_loss: 4.1369, valid_loss: 3.2072

Epoch [529/3000], train_loss: 4.0789, valid_loss: 3.4228

Epoch [530/3000], train_loss: 4.2543, valid_loss: 3.6048

Epoch [531/3000], train_loss: 4.3495, valid_loss: 3.0345

Epoch [532/3000], train_loss: 4.2354, valid_loss: 3.4482

Epoch [533/3000], train_loss: 4.1984, valid_loss: 3.5241

Epoch [534/3000], train_loss: 4.2205, valid_loss: 3.4824

Epoch [535/3000], train_loss: 4.0939, valid_loss: 3.3472

Epoch [536/3000], train_loss: 3.6870, valid_loss: 3.6086

Epoch [537/3000], train_loss: 3.9134, valid_loss: 2.7410
Save model at epoch 537 with loss 2.7410

Epoch [538/3000], train_loss: 3.5313, valid_loss: 3.1787

Epoch [539/3000], train_loss: 3.6616, valid_loss: 3.0456

Epoch [540/3000], train_loss: 3.9273, valid_loss: 3.0952

Epoch [541/3000], train_loss: 4.1920, valid_loss: 2.7938

Epoch [542/3000], train_loss: 3.8291, valid_loss: 3.1703

Epoch [543/3000], train_loss: 3.9260, valid_loss: 3.1691

Epoch [544/3000], train_loss: 3.7631, valid_loss: 2.7369
Save model at epoch 544 with loss 2.7369

Epoch [545/3000], train_loss: 3.8491, valid_loss: 3.2395

Epoch [546/3000], train_loss: 3.7851, valid_loss: 3.7860

Epoch [547/3000], train_loss: 3.6955, valid_loss: 3.1433

Epoch [548/3000], train_loss: 4.2189, valid_loss: 3.2975

Epoch [549/3000], train_loss: 4.2481, valid_loss: 3.1775

Epoch [550/3000], train_loss: 3.5630, valid_loss: 2.9807

Epoch [551/3000], train_loss: 3.7135, valid_loss: 3.2221

Epoch [552/3000], train_loss: 3.6909, valid_loss: 2.7278
Save model at epoch 552 with loss 2.7278

Epoch [553/3000], train_loss: 3.6474, valid_loss: 3.5093

Epoch [554/3000], train_loss: 4.0756, valid_loss: 3.5200

Epoch [555/3000], train_loss: 3.4822, valid_loss: 2.8081

Epoch [556/3000], train_loss: 3.6754, valid_loss: 3.0489

Epoch [557/3000], train_loss: 3.7849, valid_loss: 2.9298

Epoch [558/3000], train_loss: 4.0913, valid_loss: 2.7984

Epoch [559/3000], train_loss: 3.3838, valid_loss: 3.1054

Epoch [560/3000], train_loss: 3.8259, valid_loss: 2.6267
Save model at epoch 560 with loss 2.6267

Epoch [561/3000], train_loss: 3.6827, valid_loss: 2.7430

Epoch [562/3000], train_loss: 3.7462, valid_loss: 2.9954

Epoch [563/3000], train_loss: 3.5902, valid_loss: 3.1458

Epoch [564/3000], train_loss: 3.8507, valid_loss: 2.6183
Save model at epoch 564 with loss 2.6183

Epoch [565/3000], train_loss: 3.7270, valid_loss: 2.4635
Save model at epoch 565 with loss 2.4635

Epoch [566/3000], train_loss: 3.8513, valid_loss: 2.4863

Epoch [567/3000], train_loss: 3.4641, valid_loss: 2.3847
Save model at epoch 567 with loss 2.3847

Epoch [568/3000], train_loss: 3.7649, valid_loss: 3.6550

Epoch [569/3000], train_loss: 3.4937, valid_loss: 2.6997

Epoch [570/3000], train_loss: 3.7460, valid_loss: 2.6553

Epoch [571/3000], train_loss: 3.5043, valid_loss: 2.4702

Epoch [572/3000], train_loss: 3.4464, valid_loss: 2.4500

Epoch [573/3000], train_loss: 3.4205, valid_loss: 2.6908

Epoch [574/3000], train_loss: 3.3848, valid_loss: 2.7786

Epoch [575/3000], train_loss: 3.3991, valid_loss: 2.6016

Epoch [576/3000], train_loss: 3.4104, valid_loss: 2.9526

Epoch [577/3000], train_loss: 3.7217, valid_loss: 2.9444

Epoch [578/3000], train_loss: 3.4974, valid_loss: 3.0084

Epoch [579/3000], train_loss: 3.6424, valid_loss: 2.4543

Epoch [580/3000], train_loss: 3.6057, valid_loss: 2.7147

Epoch [581/3000], train_loss: 3.5504, valid_loss: 2.8464

Epoch [582/3000], train_loss: 3.4568, valid_loss: 2.8279

Epoch [583/3000], train_loss: 3.8781, valid_loss: 2.4673

Epoch [584/3000], train_loss: 3.3715, valid_loss: 2.7339

Epoch [585/3000], train_loss: 3.5452, valid_loss: 2.5371

Epoch [586/3000], train_loss: 3.6846, valid_loss: 2.5217

Epoch [587/3000], train_loss: 3.4544, valid_loss: 2.5521

Epoch [588/3000], train_loss: 3.6317, valid_loss: 2.3845
Save model at epoch 588 with loss 2.3845

Epoch [589/3000], train_loss: 3.5252, valid_loss: 2.3454
Save model at epoch 589 with loss 2.3454

Epoch [590/3000], train_loss: 3.5651, valid_loss: 2.5682

Epoch [591/3000], train_loss: 3.3035, valid_loss: 2.6118

Epoch [592/3000], train_loss: 3.3915, valid_loss: 2.2140
Save model at epoch 592 with loss 2.2140

Epoch [593/3000], train_loss: 3.4665, valid_loss: 2.8103

Epoch [594/3000], train_loss: 3.2082, valid_loss: 2.5197

Epoch [595/3000], train_loss: 3.4256, valid_loss: 2.2972

Epoch [596/3000], train_loss: 3.7026, valid_loss: 2.5857

Epoch [597/3000], train_loss: 3.4043, valid_loss: 2.8537

Epoch [598/3000], train_loss: 3.6070, valid_loss: 2.5649

Epoch [599/3000], train_loss: 3.2713, valid_loss: 2.5270

Epoch [600/3000], train_loss: 3.3736, valid_loss: 2.8518

Epoch [601/3000], train_loss: 3.5619, valid_loss: 2.6703

Epoch [602/3000], train_loss: 3.4932, valid_loss: 2.3752

Epoch [603/3000], train_loss: 3.4888, valid_loss: 2.4790

Epoch [604/3000], train_loss: 3.0665, valid_loss: 2.4543

Epoch [605/3000], train_loss: 3.2808, valid_loss: 2.6313

Epoch [606/3000], train_loss: 3.3446, valid_loss: 2.3982

Epoch [607/3000], train_loss: 3.5765, valid_loss: 2.7046

Epoch [608/3000], train_loss: 3.8675, valid_loss: 2.2377

Epoch [609/3000], train_loss: 3.4639, valid_loss: 2.2851

Epoch [610/3000], train_loss: 3.7356, valid_loss: 2.4351

Epoch [611/3000], train_loss: 3.4773, valid_loss: 2.2484

Epoch [612/3000], train_loss: 3.3169, valid_loss: 2.1422
Save model at epoch 612 with loss 2.1422

Epoch [613/3000], train_loss: 3.1594, valid_loss: 2.4079

Epoch [614/3000], train_loss: 3.2248, valid_loss: 2.2393

Epoch [615/3000], train_loss: 3.3373, valid_loss: 2.2485

Epoch [616/3000], train_loss: 3.2391, valid_loss: 2.5508

Epoch [617/3000], train_loss: 3.5476, valid_loss: 2.1843

Epoch [618/3000], train_loss: 2.9884, valid_loss: 2.3729

Epoch [619/3000], train_loss: 3.3507, valid_loss: 2.1228
Save model at epoch 619 with loss 2.1228

Epoch [620/3000], train_loss: 3.3320, valid_loss: 2.3701

Epoch [621/3000], train_loss: 3.3206, valid_loss: 3.3088

Epoch [622/3000], train_loss: 3.2415, valid_loss: 2.6153

Epoch [623/3000], train_loss: 3.8212, valid_loss: 2.2051

Epoch [624/3000], train_loss: 3.2328, valid_loss: 2.6790

Epoch [625/3000], train_loss: 3.2374, valid_loss: 3.2584

Epoch [626/3000], train_loss: 3.0588, valid_loss: 2.1973

Epoch [627/3000], train_loss: 2.9911, valid_loss: 2.2250

Epoch [628/3000], train_loss: 3.3597, valid_loss: 2.3268

Epoch [629/3000], train_loss: 3.1602, valid_loss: 2.0824
Save model at epoch 629 with loss 2.0824

Epoch [630/3000], train_loss: 3.6157, valid_loss: 2.0596
Save model at epoch 630 with loss 2.0596

Epoch [631/3000], train_loss: 3.0749, valid_loss: 2.5412

Epoch [632/3000], train_loss: 3.2250, valid_loss: 2.5216

Epoch [633/3000], train_loss: 3.3260, valid_loss: 2.3777

Epoch [634/3000], train_loss: 3.6514, valid_loss: 2.0663

Epoch [635/3000], train_loss: 3.6888, valid_loss: 2.3105

Epoch [636/3000], train_loss: 3.3425, valid_loss: 2.0914

Epoch [637/3000], train_loss: 3.5550, valid_loss: 2.5149

Epoch [638/3000], train_loss: 3.0380, valid_loss: 2.6412

Epoch [639/3000], train_loss: 3.3545, valid_loss: 2.5079

Epoch [640/3000], train_loss: 3.8564, valid_loss: 2.0473
Save model at epoch 640 with loss 2.0473

Epoch [641/3000], train_loss: 3.2090, valid_loss: 2.2712

Epoch [642/3000], train_loss: 3.2078, valid_loss: 2.0286
Save model at epoch 642 with loss 2.0286

Epoch [643/3000], train_loss: 3.1505, valid_loss: 2.6372

Epoch [644/3000], train_loss: 3.0053, valid_loss: 2.7705

Epoch [645/3000], train_loss: 3.4688, valid_loss: 2.0387

Epoch [646/3000], train_loss: 3.2146, valid_loss: 2.1626

Epoch [647/3000], train_loss: 3.1349, valid_loss: 2.1299

Epoch [648/3000], train_loss: 3.5413, valid_loss: 1.9951
Save model at epoch 648 with loss 1.9951

Epoch [649/3000], train_loss: 3.3056, valid_loss: 2.2592

Epoch [650/3000], train_loss: 3.2960, valid_loss: 1.8571
Save model at epoch 650 with loss 1.8571

Epoch [651/3000], train_loss: 2.9616, valid_loss: 2.1324

Epoch [652/3000], train_loss: 3.1131, valid_loss: 2.8273

Epoch [653/3000], train_loss: 3.2308, valid_loss: 2.3826

Epoch [654/3000], train_loss: 3.1060, valid_loss: 2.4532

Epoch [655/3000], train_loss: 2.9911, valid_loss: 2.1808

Epoch [656/3000], train_loss: 3.0181, valid_loss: 2.1323

Epoch [657/3000], train_loss: 3.2514, valid_loss: 2.1350

Epoch [658/3000], train_loss: 3.1513, valid_loss: 2.1919

Epoch [659/3000], train_loss: 2.9643, valid_loss: 2.3473

Epoch [660/3000], train_loss: 3.2161, valid_loss: 2.3760

Epoch [661/3000], train_loss: 3.4850, valid_loss: 2.2103

Epoch [662/3000], train_loss: 3.0287, valid_loss: 1.7879
Save model at epoch 662 with loss 1.7879

Epoch [663/3000], train_loss: 3.0975, valid_loss: 2.1409

Epoch [664/3000], train_loss: 3.1418, valid_loss: 2.2395

Epoch [665/3000], train_loss: 3.4812, valid_loss: 2.1949

Epoch [666/3000], train_loss: 3.3264, valid_loss: 2.0730

Epoch [667/3000], train_loss: 3.4318, valid_loss: 2.0887

Epoch [668/3000], train_loss: 3.2469, valid_loss: 1.9538

Epoch [669/3000], train_loss: 3.1776, valid_loss: 2.1768

Epoch [670/3000], train_loss: 3.4656, valid_loss: 2.5195

Epoch [671/3000], train_loss: 2.9943, valid_loss: 2.4381

Epoch [672/3000], train_loss: 3.2986, valid_loss: 2.3256

Epoch [673/3000], train_loss: 3.2104, valid_loss: 2.4751

Epoch [674/3000], train_loss: 3.1899, valid_loss: 2.0252

Epoch [675/3000], train_loss: 3.4970, valid_loss: 2.0737

Epoch [676/3000], train_loss: 3.2773, valid_loss: 2.0069

Epoch [677/3000], train_loss: 3.1178, valid_loss: 1.8919

Epoch [678/3000], train_loss: 3.1930, valid_loss: 2.5538

Epoch [679/3000], train_loss: 2.9576, valid_loss: 1.7897

Epoch [680/3000], train_loss: 3.1907, valid_loss: 2.5404

Epoch [681/3000], train_loss: 3.1159, valid_loss: 2.1034

Epoch [682/3000], train_loss: 2.9343, valid_loss: 2.0956

Epoch [683/3000], train_loss: 3.2754, valid_loss: 2.1718

Epoch [684/3000], train_loss: 3.0902, valid_loss: 2.0416

Epoch [685/3000], train_loss: 3.3043, valid_loss: 2.0294

Epoch [686/3000], train_loss: 2.9668, valid_loss: 2.1526

Epoch [687/3000], train_loss: 3.0913, valid_loss: 2.1789

Epoch [688/3000], train_loss: 3.0584, valid_loss: 2.2747

Epoch [689/3000], train_loss: 3.1772, valid_loss: 1.9139

Epoch [690/3000], train_loss: 2.8891, valid_loss: 2.1570

Epoch [691/3000], train_loss: 3.0982, valid_loss: 2.1705

Epoch [692/3000], train_loss: 3.1729, valid_loss: 2.3773

Epoch [693/3000], train_loss: 3.2241, valid_loss: 1.9952

Epoch [694/3000], train_loss: 3.0129, valid_loss: 1.7508
Save model at epoch 694 with loss 1.7508

Epoch [695/3000], train_loss: 2.8278, valid_loss: 2.3692

Epoch [696/3000], train_loss: 3.2394, valid_loss: 2.4467

Epoch [697/3000], train_loss: 3.0604, valid_loss: 2.1927

Epoch [698/3000], train_loss: 2.9349, valid_loss: 1.7927

Epoch [699/3000], train_loss: 3.0609, valid_loss: 2.3340

Epoch [700/3000], train_loss: 3.0617, valid_loss: 2.4609

Epoch [701/3000], train_loss: 3.0820, valid_loss: 1.9945

Epoch [702/3000], train_loss: 3.0981, valid_loss: 2.2572

Epoch [703/3000], train_loss: 3.0606, valid_loss: 2.1414

Epoch [704/3000], train_loss: 3.2532, valid_loss: 2.6410

Epoch [705/3000], train_loss: 3.2186, valid_loss: 2.2118

Epoch [706/3000], train_loss: 2.9537, valid_loss: 2.2769

Epoch [707/3000], train_loss: 2.9943, valid_loss: 1.9269

Epoch [708/3000], train_loss: 2.8406, valid_loss: 2.2452

Epoch [709/3000], train_loss: 3.3586, valid_loss: 2.0501

Epoch [710/3000], train_loss: 2.9238, valid_loss: 2.1393

Epoch [711/3000], train_loss: 3.1225, valid_loss: 1.9631

Epoch [712/3000], train_loss: 3.2096, valid_loss: 2.2211

Epoch [713/3000], train_loss: 2.8889, valid_loss: 2.4276

Epoch [714/3000], train_loss: 3.2083, valid_loss: 1.9814

Epoch [715/3000], train_loss: 3.3679, valid_loss: 1.8102

Epoch [716/3000], train_loss: 3.0360, valid_loss: 2.2034

Epoch [717/3000], train_loss: 3.0696, valid_loss: 2.0536

Epoch [718/3000], train_loss: 3.0861, valid_loss: 2.0861

Epoch [719/3000], train_loss: 3.2688, valid_loss: 2.1126

Epoch [720/3000], train_loss: 3.0740, valid_loss: 2.1819

Epoch [721/3000], train_loss: 3.0979, valid_loss: 2.0717

Epoch [722/3000], train_loss: 3.4194, valid_loss: 1.9959

Epoch [723/3000], train_loss: 3.5010, valid_loss: 2.5011

Epoch [724/3000], train_loss: 3.0305, valid_loss: 1.8650

Epoch [725/3000], train_loss: 3.0661, valid_loss: 2.3308

Epoch [726/3000], train_loss: 3.2833, valid_loss: 1.9913

Epoch [727/3000], train_loss: 3.3794, valid_loss: 1.7699

Epoch [728/3000], train_loss: 3.2049, valid_loss: 1.7127
Save model at epoch 728 with loss 1.7127

Epoch [729/3000], train_loss: 2.7646, valid_loss: 2.0044

Epoch [730/3000], train_loss: 3.0360, valid_loss: 2.0170

Epoch [731/3000], train_loss: 2.9898, valid_loss: 2.1985

Epoch [732/3000], train_loss: 3.1608, valid_loss: 2.1170

Epoch [733/3000], train_loss: 3.0732, valid_loss: 2.0093

Epoch [734/3000], train_loss: 3.0239, valid_loss: 2.2727

Epoch [735/3000], train_loss: 3.1499, valid_loss: 2.4081

Epoch [736/3000], train_loss: 2.6601, valid_loss: 1.8174

Epoch [737/3000], train_loss: 2.9317, valid_loss: 1.9916

Epoch [738/3000], train_loss: 3.1715, valid_loss: 1.9243

Epoch [739/3000], train_loss: 3.1036, valid_loss: 1.8895

Epoch [740/3000], train_loss: 2.8464, valid_loss: 2.1792

Epoch [741/3000], train_loss: 2.7999, valid_loss: 2.1123

Epoch [742/3000], train_loss: 2.8521, valid_loss: 1.9708

Epoch [743/3000], train_loss: 2.7188, valid_loss: 2.1582

Epoch [744/3000], train_loss: 2.9676, valid_loss: 2.1271

Epoch [745/3000], train_loss: 2.9328, valid_loss: 2.2306

Epoch [746/3000], train_loss: 3.0787, valid_loss: 1.9876

Epoch [747/3000], train_loss: 2.9481, valid_loss: 1.8868

Epoch [748/3000], train_loss: 2.8066, valid_loss: 2.7573

Epoch [749/3000], train_loss: 2.8682, valid_loss: 1.7966

Epoch [750/3000], train_loss: 2.9193, valid_loss: 2.2713

Epoch [751/3000], train_loss: 2.8022, valid_loss: 1.9947

Epoch [752/3000], train_loss: 2.7900, valid_loss: 2.1455

Epoch [753/3000], train_loss: 2.8028, valid_loss: 1.7591

Epoch [754/3000], train_loss: 3.1209, valid_loss: 2.0331

Epoch [755/3000], train_loss: 2.9065, valid_loss: 2.1083

Epoch [756/3000], train_loss: 2.7693, valid_loss: 2.2994

Epoch [757/3000], train_loss: 2.8298, valid_loss: 1.9148

Epoch [758/3000], train_loss: 2.9979, valid_loss: 1.7297

Epoch [759/3000], train_loss: 2.6623, valid_loss: 1.9116

Epoch [760/3000], train_loss: 2.7965, valid_loss: 1.9375

Epoch [761/3000], train_loss: 2.9098, valid_loss: 2.0072

Epoch [762/3000], train_loss: 2.5983, valid_loss: 1.7083
Save model at epoch 762 with loss 1.7083

Epoch [763/3000], train_loss: 2.9003, valid_loss: 1.8300

Epoch [764/3000], train_loss: 3.0624, valid_loss: 1.7692

Epoch [765/3000], train_loss: 2.9578, valid_loss: 1.6006
Save model at epoch 765 with loss 1.6006

Epoch [766/3000], train_loss: 3.0352, valid_loss: 2.1146

Epoch [767/3000], train_loss: 2.7599, valid_loss: 1.9688

Epoch [768/3000], train_loss: 3.0910, valid_loss: 2.0876

Epoch [769/3000], train_loss: 3.3707, valid_loss: 1.9991

Epoch [770/3000], train_loss: 2.8658, valid_loss: 1.7787

Epoch [771/3000], train_loss: 2.8506, valid_loss: 1.9591

Epoch [772/3000], train_loss: 2.9500, valid_loss: 2.0627

Epoch [773/3000], train_loss: 2.8882, valid_loss: 2.0067

Epoch [774/3000], train_loss: 2.8756, valid_loss: 2.3541

Epoch [775/3000], train_loss: 2.7428, valid_loss: 2.2014

Epoch [776/3000], train_loss: 2.9844, valid_loss: 2.0609

Epoch [777/3000], train_loss: 3.6928, valid_loss: 1.7654

Epoch [778/3000], train_loss: 2.9956, valid_loss: 1.9002

Epoch [779/3000], train_loss: 3.0246, valid_loss: 2.2217

Epoch [780/3000], train_loss: 3.1674, valid_loss: 1.9862

Epoch [781/3000], train_loss: 2.7531, valid_loss: 2.1668

Epoch [782/3000], train_loss: 3.4322, valid_loss: 1.6863

Epoch [783/3000], train_loss: 2.9388, valid_loss: 1.6861

Epoch [784/3000], train_loss: 2.8418, valid_loss: 2.0047

Epoch [785/3000], train_loss: 2.8815, valid_loss: 2.1096

Epoch [786/3000], train_loss: 2.8173, valid_loss: 2.0582

Epoch [787/3000], train_loss: 2.8957, valid_loss: 2.1171

Epoch [788/3000], train_loss: 2.8688, valid_loss: 2.0376

Epoch [789/3000], train_loss: 2.7699, valid_loss: 1.7625

Epoch [790/3000], train_loss: 3.0395, valid_loss: 1.7495

Epoch [791/3000], train_loss: 2.9444, valid_loss: 1.9365

Epoch [792/3000], train_loss: 2.9245, valid_loss: 2.2395

Epoch [793/3000], train_loss: 3.0600, valid_loss: 2.1133

Epoch [794/3000], train_loss: 3.0679, valid_loss: 2.1440

Epoch [795/3000], train_loss: 2.9589, valid_loss: 1.7671

Epoch [796/3000], train_loss: 2.8306, valid_loss: 1.8612

Epoch [797/3000], train_loss: 2.5431, valid_loss: 1.9030

Epoch [798/3000], train_loss: 3.1248, valid_loss: 2.0337

Epoch [799/3000], train_loss: 2.6442, valid_loss: 2.0134

Epoch [800/3000], train_loss: 2.8243, valid_loss: 2.0259

Epoch [801/3000], train_loss: 3.1326, valid_loss: 1.9633

Epoch [802/3000], train_loss: 2.8724, valid_loss: 1.9568

Epoch [803/3000], train_loss: 2.7108, valid_loss: 1.9343

Epoch [804/3000], train_loss: 2.6563, valid_loss: 2.0751

Epoch [805/3000], train_loss: 2.7243, valid_loss: 1.7941

Epoch [806/3000], train_loss: 2.5082, valid_loss: 1.7068

Epoch [807/3000], train_loss: 2.8638, valid_loss: 2.3495

Epoch [808/3000], train_loss: 3.1135, valid_loss: 1.7045

Epoch [809/3000], train_loss: 3.0150, valid_loss: 1.7873

Epoch [810/3000], train_loss: 3.1601, valid_loss: 1.8737

Epoch [811/3000], train_loss: 3.1622, valid_loss: 1.7745

Epoch [812/3000], train_loss: 3.1676, valid_loss: 2.5977

Epoch [813/3000], train_loss: 3.3860, valid_loss: 2.3975

Epoch [814/3000], train_loss: 2.7338, valid_loss: 1.9638

Epoch [815/3000], train_loss: 2.7483, valid_loss: 1.6785

Epoch [816/3000], train_loss: 3.0028, valid_loss: 2.3197

Epoch [817/3000], train_loss: 2.6658, valid_loss: 1.7499

Epoch [818/3000], train_loss: 2.8284, valid_loss: 2.0768

Epoch [819/3000], train_loss: 2.7713, valid_loss: 2.3525

Epoch [820/3000], train_loss: 2.9004, valid_loss: 1.7777

Epoch [821/3000], train_loss: 3.1012, valid_loss: 2.0039

Epoch [822/3000], train_loss: 2.8480, valid_loss: 1.7974

Epoch [823/3000], train_loss: 2.9483, valid_loss: 1.9230

Epoch [824/3000], train_loss: 2.8396, valid_loss: 2.2969

Epoch [825/3000], train_loss: 2.8919, valid_loss: 1.7791

Epoch [826/3000], train_loss: 2.8558, valid_loss: 2.0460

Epoch [827/3000], train_loss: 2.7187, valid_loss: 1.7774

Epoch [828/3000], train_loss: 2.5892, valid_loss: 2.0487

Epoch [829/3000], train_loss: 2.9972, valid_loss: 1.9055

Epoch [830/3000], train_loss: 2.8294, valid_loss: 1.8232

Epoch [831/3000], train_loss: 2.7132, valid_loss: 1.9399

Epoch [832/3000], train_loss: 2.5761, valid_loss: 2.1424

Epoch [833/3000], train_loss: 2.7093, valid_loss: 1.7393

Epoch [834/3000], train_loss: 2.9523, valid_loss: 2.0827

Epoch [835/3000], train_loss: 3.0623, valid_loss: 1.8772

Epoch [836/3000], train_loss: 2.7254, valid_loss: 1.9470

Epoch [837/3000], train_loss: 3.1114, valid_loss: 1.8397

Epoch [838/3000], train_loss: 3.2351, valid_loss: 1.8934

Epoch [839/3000], train_loss: 2.8959, valid_loss: 1.5943
Save model at epoch 839 with loss 1.5943

Epoch [840/3000], train_loss: 2.6434, valid_loss: 2.0318

Epoch [841/3000], train_loss: 2.8406, valid_loss: 1.8353

Epoch [842/3000], train_loss: 2.8297, valid_loss: 1.7469

Epoch [843/3000], train_loss: 2.8733, valid_loss: 1.8632

Epoch [844/3000], train_loss: 2.7448, valid_loss: 1.9659

Epoch [845/3000], train_loss: 2.6968, valid_loss: 2.0698

Epoch [846/3000], train_loss: 2.7731, valid_loss: 1.7212

Epoch [847/3000], train_loss: 3.4105, valid_loss: 1.8117

Epoch [848/3000], train_loss: 2.9925, valid_loss: 1.9717

Epoch [849/3000], train_loss: 2.9365, valid_loss: 1.6845

Epoch [850/3000], train_loss: 2.7046, valid_loss: 1.8057

Epoch [851/3000], train_loss: 2.5402, valid_loss: 1.7218

Epoch [852/3000], train_loss: 3.0923, valid_loss: 1.9034

Epoch [853/3000], train_loss: 2.6137, valid_loss: 1.6721

Epoch [854/3000], train_loss: 2.9592, valid_loss: 1.9021

Epoch [855/3000], train_loss: 2.6539, valid_loss: 1.8542

Epoch [856/3000], train_loss: 2.8199, valid_loss: 1.8997

Epoch [857/3000], train_loss: 2.7415, valid_loss: 2.1036

Epoch [858/3000], train_loss: 2.7749, valid_loss: 1.6324

Epoch [859/3000], train_loss: 2.9600, valid_loss: 1.6774

Epoch [860/3000], train_loss: 2.9705, valid_loss: 1.6057

Epoch [861/3000], train_loss: 2.8669, valid_loss: 1.9593

Epoch [862/3000], train_loss: 2.8817, valid_loss: 1.6698

Epoch [863/3000], train_loss: 2.9690, valid_loss: 2.0565

Epoch [864/3000], train_loss: 3.0164, valid_loss: 1.8415

Epoch [865/3000], train_loss: 2.5527, valid_loss: 2.4257

Epoch [866/3000], train_loss: 2.7556, valid_loss: 2.3173

Epoch [867/3000], train_loss: 2.8646, valid_loss: 2.1544

Epoch [868/3000], train_loss: 2.9385, valid_loss: 1.7399

Epoch [869/3000], train_loss: 2.9253, valid_loss: 1.8537

Epoch [870/3000], train_loss: 2.8127, valid_loss: 2.3991

Epoch [871/3000], train_loss: 3.1391, valid_loss: 2.1131

Epoch [872/3000], train_loss: 2.7605, valid_loss: 2.0618

Epoch [873/3000], train_loss: 2.6885, valid_loss: 1.8554

Epoch [874/3000], train_loss: 2.7160, valid_loss: 1.6266

Epoch [875/3000], train_loss: 2.7690, valid_loss: 1.7781

Epoch [876/3000], train_loss: 2.6789, valid_loss: 1.4719
Save model at epoch 876 with loss 1.4719

Epoch [877/3000], train_loss: 2.7553, valid_loss: 2.0571

Epoch [878/3000], train_loss: 2.6096, valid_loss: 2.0199

Epoch [879/3000], train_loss: 2.8228, valid_loss: 1.9825

Epoch [880/3000], train_loss: 2.7985, valid_loss: 1.8358

Epoch [881/3000], train_loss: 2.5873, valid_loss: 1.9522

Epoch [882/3000], train_loss: 2.7100, valid_loss: 2.0465

Epoch [883/3000], train_loss: 2.9439, valid_loss: 2.0861

Epoch [884/3000], train_loss: 2.9150, valid_loss: 1.5882

Epoch [885/3000], train_loss: 2.5782, valid_loss: 1.6226

Epoch [886/3000], train_loss: 2.5598, valid_loss: 1.6140

Epoch [887/3000], train_loss: 2.8279, valid_loss: 1.8191

Epoch [888/3000], train_loss: 2.7219, valid_loss: 1.8977

Epoch [889/3000], train_loss: 2.8254, valid_loss: 1.8459

Epoch [890/3000], train_loss: 2.7000, valid_loss: 1.9068

Epoch [891/3000], train_loss: 2.6990, valid_loss: 1.6924

Epoch [892/3000], train_loss: 3.0504, valid_loss: 2.0172

Epoch [893/3000], train_loss: 2.7243, valid_loss: 1.8779

Epoch [894/3000], train_loss: 2.6046, valid_loss: 1.7484

Epoch [895/3000], train_loss: 2.7857, valid_loss: 1.8133

Epoch [896/3000], train_loss: 2.7650, valid_loss: 1.7723

Epoch [897/3000], train_loss: 2.3514, valid_loss: 1.7914

Epoch [898/3000], train_loss: 2.6602, valid_loss: 1.8272

Epoch [899/3000], train_loss: 2.5062, valid_loss: 2.0085

Epoch [900/3000], train_loss: 2.9052, valid_loss: 1.8526

Epoch [901/3000], train_loss: 2.7773, valid_loss: 1.7352

Epoch [902/3000], train_loss: 2.9514, valid_loss: 1.6270

Epoch [903/3000], train_loss: 2.6669, valid_loss: 1.7481

Epoch [904/3000], train_loss: 2.6936, valid_loss: 1.6291

Epoch [905/3000], train_loss: 3.0233, valid_loss: 1.6073

Epoch [906/3000], train_loss: 2.6689, valid_loss: 1.7919

Epoch [907/3000], train_loss: 2.9068, valid_loss: 1.7451

Epoch [908/3000], train_loss: 3.1664, valid_loss: 1.8228

Epoch [909/3000], train_loss: 2.4853, valid_loss: 1.8639

Epoch [910/3000], train_loss: 2.6953, valid_loss: 1.7289

Epoch [911/3000], train_loss: 2.5035, valid_loss: 2.4439

Epoch [912/3000], train_loss: 2.5653, valid_loss: 2.2329

Epoch [913/3000], train_loss: 2.5754, valid_loss: 1.7671

Epoch [914/3000], train_loss: 2.8915, valid_loss: 1.9111

Epoch [915/3000], train_loss: 2.7131, valid_loss: 1.8942

Epoch [916/3000], train_loss: 2.6616, valid_loss: 2.0553

Epoch [917/3000], train_loss: 3.0364, valid_loss: 1.6312

Epoch [918/3000], train_loss: 2.5390, valid_loss: 1.6905

Epoch [919/3000], train_loss: 2.5676, valid_loss: 1.7972

Epoch [920/3000], train_loss: 2.7620, valid_loss: 1.5043

Epoch [921/3000], train_loss: 2.6554, valid_loss: 1.6744

Epoch [922/3000], train_loss: 2.8193, valid_loss: 1.6290

Epoch [923/3000], train_loss: 2.5769, valid_loss: 1.7288

Epoch [924/3000], train_loss: 2.6746, valid_loss: 1.8560

Epoch [925/3000], train_loss: 2.7063, valid_loss: 1.8942

Epoch [926/3000], train_loss: 2.6259, valid_loss: 1.5624

Epoch [927/3000], train_loss: 2.8243, valid_loss: 1.9223

Epoch [928/3000], train_loss: 2.4861, valid_loss: 1.9118

Epoch [929/3000], train_loss: 2.7686, valid_loss: 1.6080

Epoch [930/3000], train_loss: 2.7746, valid_loss: 1.8282

Epoch [931/3000], train_loss: 2.5395, valid_loss: 1.7165

Epoch [932/3000], train_loss: 2.5902, valid_loss: 1.7103

Epoch [933/3000], train_loss: 2.4832, valid_loss: 1.5418

Epoch [934/3000], train_loss: 2.6602, valid_loss: 1.9813

Epoch [935/3000], train_loss: 2.6388, valid_loss: 1.7738

Epoch [936/3000], train_loss: 2.7161, valid_loss: 1.7905

Epoch [937/3000], train_loss: 2.9119, valid_loss: 1.4895

Epoch [938/3000], train_loss: 2.4093, valid_loss: 1.9842

Epoch [939/3000], train_loss: 2.6296, valid_loss: 1.6858

Epoch [940/3000], train_loss: 2.5469, valid_loss: 1.7075

Epoch [941/3000], train_loss: 2.8520, valid_loss: 1.6882

Epoch [942/3000], train_loss: 2.8345, valid_loss: 1.6951

Epoch [943/3000], train_loss: 2.3829, valid_loss: 1.6220

Epoch [944/3000], train_loss: 2.7467, valid_loss: 2.1909

Epoch [945/3000], train_loss: 2.8455, valid_loss: 1.8138

Epoch [946/3000], train_loss: 2.7298, valid_loss: 1.6537

Epoch [947/3000], train_loss: 2.7366, valid_loss: 1.6312

Epoch [948/3000], train_loss: 3.0015, valid_loss: 2.2575

Epoch [949/3000], train_loss: 2.4558, valid_loss: 2.0268

Epoch [950/3000], train_loss: 2.4970, valid_loss: 2.0233

Epoch [951/3000], train_loss: 2.6654, valid_loss: 1.6513

Epoch [952/3000], train_loss: 2.6947, valid_loss: 2.5442

Epoch [953/3000], train_loss: 2.9091, valid_loss: 2.0382

Epoch [954/3000], train_loss: 2.5025, valid_loss: 2.2741

Epoch [955/3000], train_loss: 2.5474, valid_loss: 1.6417

Epoch [956/3000], train_loss: 2.5242, valid_loss: 1.5623

Epoch [957/3000], train_loss: 2.6382, valid_loss: 1.4995

Epoch [958/3000], train_loss: 2.5268, valid_loss: 1.7880

Epoch [959/3000], train_loss: 2.5506, valid_loss: 1.6008

Epoch [960/3000], train_loss: 2.7840, valid_loss: 1.7673

Epoch [961/3000], train_loss: 2.5662, valid_loss: 1.7214

Epoch [962/3000], train_loss: 2.6321, valid_loss: 1.8616

Epoch [963/3000], train_loss: 2.6821, valid_loss: 1.8044

Epoch [964/3000], train_loss: 2.4854, valid_loss: 1.6661

Epoch [965/3000], train_loss: 2.6682, valid_loss: 1.6460

Epoch [966/3000], train_loss: 2.6291, valid_loss: 1.6828

Epoch [967/3000], train_loss: 2.4106, valid_loss: 1.8688

Epoch [968/3000], train_loss: 2.7001, valid_loss: 1.7307

Epoch [969/3000], train_loss: 2.5665, valid_loss: 1.9206

Epoch [970/3000], train_loss: 2.7048, valid_loss: 1.6476

Epoch [971/3000], train_loss: 2.6443, valid_loss: 2.1188

Epoch [972/3000], train_loss: 2.4975, valid_loss: 1.5846

Epoch [973/3000], train_loss: 2.5683, valid_loss: 1.8202

Epoch [974/3000], train_loss: 2.5021, valid_loss: 1.7818

Epoch [975/3000], train_loss: 2.7639, valid_loss: 1.8854

Epoch [976/3000], train_loss: 2.6359, valid_loss: 1.5399

Epoch [977/3000], train_loss: 2.4292, valid_loss: 1.5342

Epoch [978/3000], train_loss: 2.8944, valid_loss: 1.7094

Epoch [979/3000], train_loss: 2.4655, valid_loss: 1.6083

Epoch [980/3000], train_loss: 2.3855, valid_loss: 1.6820

Epoch [981/3000], train_loss: 2.7111, valid_loss: 1.5511

Epoch [982/3000], train_loss: 2.5984, valid_loss: 2.0117

Epoch [983/3000], train_loss: 2.6484, valid_loss: 1.5289

Epoch [984/3000], train_loss: 2.8065, valid_loss: 1.8859

Epoch [985/3000], train_loss: 2.4273, valid_loss: 1.5166

Epoch [986/3000], train_loss: 2.3929, valid_loss: 1.4917

Epoch [987/3000], train_loss: 2.6262, valid_loss: 1.6764

Epoch [988/3000], train_loss: 2.6079, valid_loss: 1.8337

Epoch [989/3000], train_loss: 2.5498, valid_loss: 1.8667

Epoch [990/3000], train_loss: 2.5625, valid_loss: 1.7432

Epoch [991/3000], train_loss: 2.5158, valid_loss: 1.8663

Epoch [992/3000], train_loss: 2.6095, valid_loss: 1.9378

Epoch [993/3000], train_loss: 2.5431, valid_loss: 1.5607

Epoch [994/3000], train_loss: 2.4844, valid_loss: 1.6585

Epoch [995/3000], train_loss: 2.5883, valid_loss: 1.8020

Epoch [996/3000], train_loss: 2.5881, valid_loss: 1.6978

Epoch [997/3000], train_loss: 2.6670, valid_loss: 1.8035

Epoch [998/3000], train_loss: 2.5914, valid_loss: 1.6412

Epoch [999/3000], train_loss: 2.8970, valid_loss: 1.6045

Epoch [1000/3000], train_loss: 2.7678, valid_loss: 1.9208

Epoch [1001/3000], train_loss: 2.3187, valid_loss: 2.0448

Epoch [1002/3000], train_loss: 2.8953, valid_loss: 1.5305

Epoch [1003/3000], train_loss: 2.6773, valid_loss: 2.0061

Epoch [1004/3000], train_loss: 2.4002, valid_loss: 1.6195

Epoch [1005/3000], train_loss: 2.4218, valid_loss: 1.9053

Epoch [1006/3000], train_loss: 2.4133, valid_loss: 1.7474

Epoch [1007/3000], train_loss: 2.2606, valid_loss: 1.6932

Epoch [1008/3000], train_loss: 2.6239, valid_loss: 1.6943

Epoch [1009/3000], train_loss: 2.5747, valid_loss: 1.5222

Epoch [1010/3000], train_loss: 2.7750, valid_loss: 1.6663

Epoch [1011/3000], train_loss: 2.7333, valid_loss: 2.1338

Epoch [1012/3000], train_loss: 2.4980, valid_loss: 1.7705

Epoch [1013/3000], train_loss: 2.4525, valid_loss: 1.6679

Epoch [1014/3000], train_loss: 2.6077, valid_loss: 1.6234

Epoch [1015/3000], train_loss: 2.4709, valid_loss: 1.6591

Epoch [1016/3000], train_loss: 2.5662, valid_loss: 1.8046

Epoch [1017/3000], train_loss: 2.7107, valid_loss: 1.4697
Save model at epoch 1017 with loss 1.4697

Epoch [1018/3000], train_loss: 2.4732, valid_loss: 1.8847

Epoch [1019/3000], train_loss: 2.6379, valid_loss: 1.7271

Epoch [1020/3000], train_loss: 2.4473, valid_loss: 1.5415

Epoch [1021/3000], train_loss: 2.4034, valid_loss: 1.8809

Epoch [1022/3000], train_loss: 2.7248, valid_loss: 1.7204

Epoch [1023/3000], train_loss: 2.4165, valid_loss: 2.0009

Epoch [1024/3000], train_loss: 2.5759, valid_loss: 1.7257

Epoch [1025/3000], train_loss: 3.0217, valid_loss: 1.9711

Epoch [1026/3000], train_loss: 2.5870, valid_loss: 1.6082

Epoch [1027/3000], train_loss: 2.3221, valid_loss: 1.4629
Save model at epoch 1027 with loss 1.4629

Epoch [1028/3000], train_loss: 2.4899, valid_loss: 1.6729

Epoch [1029/3000], train_loss: 2.4413, valid_loss: 1.4766

Epoch [1030/3000], train_loss: 2.7231, valid_loss: 1.7432

Epoch [1031/3000], train_loss: 2.4260, valid_loss: 1.6365

Epoch [1032/3000], train_loss: 2.4788, valid_loss: 1.8087

Epoch [1033/3000], train_loss: 2.3885, valid_loss: 1.5375

Epoch [1034/3000], train_loss: 2.4362, valid_loss: 1.5431

Epoch [1035/3000], train_loss: 2.3379, valid_loss: 1.7144

Epoch [1036/3000], train_loss: 2.1804, valid_loss: 1.8132

Epoch [1037/3000], train_loss: 2.4670, valid_loss: 1.5878

Epoch [1038/3000], train_loss: 2.5724, valid_loss: 1.6719

Epoch [1039/3000], train_loss: 2.5170, valid_loss: 1.7935

Epoch [1040/3000], train_loss: 2.4093, valid_loss: 2.0688

Epoch [1041/3000], train_loss: 2.4972, valid_loss: 1.5312

Epoch [1042/3000], train_loss: 2.5321, valid_loss: 1.5172

Epoch [1043/3000], train_loss: 2.5656, valid_loss: 1.8277

Epoch [1044/3000], train_loss: 2.5866, valid_loss: 1.5704

Epoch [1045/3000], train_loss: 2.6909, valid_loss: 1.5767

Epoch [1046/3000], train_loss: 2.3787, valid_loss: 1.6415

Epoch [1047/3000], train_loss: 2.5399, valid_loss: 1.8353

Epoch [1048/3000], train_loss: 2.4641, valid_loss: 1.6913

Epoch [1049/3000], train_loss: 2.2962, valid_loss: 1.7292

Epoch [1050/3000], train_loss: 2.6402, valid_loss: 1.5328

Epoch [1051/3000], train_loss: 2.4643, valid_loss: 1.4180
Save model at epoch 1051 with loss 1.4180

Epoch [1052/3000], train_loss: 2.2673, valid_loss: 1.5501

Epoch [1053/3000], train_loss: 2.4924, valid_loss: 1.8863

Epoch [1054/3000], train_loss: 2.6539, valid_loss: 1.4037
Save model at epoch 1054 with loss 1.4037

Epoch [1055/3000], train_loss: 2.5705, valid_loss: 1.7693

Epoch [1056/3000], train_loss: 3.0047, valid_loss: 1.8522

Epoch [1057/3000], train_loss: 2.3424, valid_loss: 1.6456

Epoch [1058/3000], train_loss: 2.3456, valid_loss: 1.7291

Epoch [1059/3000], train_loss: 2.6359, valid_loss: 1.5829

Epoch [1060/3000], train_loss: 2.7155, valid_loss: 1.5257

Epoch [1061/3000], train_loss: 2.7420, valid_loss: 1.9634

Epoch [1062/3000], train_loss: 2.7076, valid_loss: 1.8178

Epoch [1063/3000], train_loss: 2.2548, valid_loss: 1.5531

Epoch [1064/3000], train_loss: 2.4989, valid_loss: 1.6830

Epoch [1065/3000], train_loss: 2.5996, valid_loss: 1.9258

Epoch [1066/3000], train_loss: 2.6149, valid_loss: 1.6556

Epoch [1067/3000], train_loss: 2.5197, valid_loss: 1.6002

Epoch [1068/3000], train_loss: 2.6779, valid_loss: 1.8396

Epoch [1069/3000], train_loss: 2.5362, valid_loss: 1.4864

Epoch [1070/3000], train_loss: 2.4745, valid_loss: 1.6274

Epoch [1071/3000], train_loss: 2.9077, valid_loss: 1.7470

Epoch [1072/3000], train_loss: 2.8817, valid_loss: 1.7404

Epoch [1073/3000], train_loss: 2.3521, valid_loss: 2.4464

Epoch [1074/3000], train_loss: 2.2164, valid_loss: 2.1596

Epoch [1075/3000], train_loss: 2.4103, valid_loss: 1.5035

Epoch [1076/3000], train_loss: 2.4451, valid_loss: 1.4730

Epoch [1077/3000], train_loss: 2.7309, valid_loss: 1.7887

Epoch [1078/3000], train_loss: 2.3130, valid_loss: 1.8988

Epoch [1079/3000], train_loss: 2.6081, valid_loss: 1.7063

Epoch [1080/3000], train_loss: 2.9178, valid_loss: 1.6305

Epoch [1081/3000], train_loss: 2.5339, valid_loss: 1.4779

Epoch [1082/3000], train_loss: 2.3276, valid_loss: 1.4887

Epoch [1083/3000], train_loss: 2.5682, valid_loss: 1.8013

Epoch [1084/3000], train_loss: 2.3729, valid_loss: 1.4600

Epoch [1248/3000], train_loss: 2.4699, valid_loss: 1.6456

Epoch [1249/3000], train_loss: 2.3295, valid_loss: 1.7310

Epoch [1250/3000], train_loss: 2.7221, valid_loss: 1.3797

Epoch [1251/3000], train_loss: 2.5044, valid_loss: 1.7545

Epoch [1252/3000], train_loss: 2.4828, valid_loss: 1.5312

Epoch [1253/3000], train_loss: 2.5347, valid_loss: 1.5977

Epoch [1254/3000], train_loss: 2.2785, valid_loss: 1.6766

Epoch [1255/3000], train_loss: 2.3821, valid_loss: 1.4178

Epoch [1256/3000], train_loss: 2.4426, valid_loss: 1.5217

Epoch [1257/3000], train_loss: 2.4005, valid_loss: 1.3000
Save model at epoch 1257 with loss 1.3000

Epoch [1258/3000], train_loss: 2.3180, valid_loss: 1.6646

Epoch [1259/3000], train_loss: 2.3955, valid_loss: 1.9645

Epoch [1260/3000], train_loss: 2.2478, valid_loss: 1.4828

Epoch [1261/3000], train_loss: 2.3232, valid_loss: 1.4400

Epoch [1262/3000], train_loss: 2.1197, valid_loss: 1.5484

Epoch [1263/3000], train_loss: 2.2262, valid_loss: 1.6222

Epoch [1264/3000], train_loss: 2.4733, valid_loss: 1.5139

Epoch [1265/3000], train_loss: 2.6760, valid_loss: 1.4488

Epoch [1266/3000], train_loss: 2.3562, valid_loss: 1.5562

Epoch [1267/3000], train_loss: 2.1178, valid_loss: 1.6972

Epoch [1268/3000], train_loss: 2.5883, valid_loss: 1.4122

Epoch [1269/3000], train_loss: 2.3318, valid_loss: 1.3823

Epoch [1270/3000], train_loss: 2.2268, valid_loss: 1.4048

Epoch [1271/3000], train_loss: 2.3451, valid_loss: 1.5284

Epoch [1272/3000], train_loss: 2.5563, valid_loss: 1.3873

Epoch [1273/3000], train_loss: 2.2820, valid_loss: 1.2849
Save model at epoch 1273 with loss 1.2849

Epoch [1274/3000], train_loss: 2.1983, valid_loss: 1.4369

Epoch [1275/3000], train_loss: 2.1109, valid_loss: 1.4053

Epoch [1276/3000], train_loss: 2.4061, valid_loss: 1.4321

Epoch [1277/3000], train_loss: 2.2141, valid_loss: 1.4103

Epoch [1278/3000], train_loss: 2.2669, valid_loss: 1.4417

Epoch [1279/3000], train_loss: 2.6216, valid_loss: 1.8534

Epoch [1280/3000], train_loss: 2.7784, valid_loss: 2.0174

Epoch [1281/3000], train_loss: 2.3862, valid_loss: 1.5803

Epoch [1282/3000], train_loss: 2.2894, valid_loss: 1.4159

Epoch [1283/3000], train_loss: 2.1670, valid_loss: 1.4633

Epoch [1284/3000], train_loss: 2.3316, valid_loss: 1.6684

Epoch [1285/3000], train_loss: 2.2178, valid_loss: 1.7489

Epoch [1286/3000], train_loss: 2.9230, valid_loss: 1.6420

Epoch [1287/3000], train_loss: 2.9281, valid_loss: 1.2823
Save model at epoch 1287 with loss 1.2823

Epoch [1288/3000], train_loss: 2.2373, valid_loss: 1.8902

Epoch [1289/3000], train_loss: 2.5489, valid_loss: 1.5497

Epoch [1290/3000], train_loss: 2.3029, valid_loss: 1.5005

Epoch [1291/3000], train_loss: 2.3688, valid_loss: 1.4703

Epoch [1292/3000], train_loss: 2.3299, valid_loss: 1.4514

Epoch [1293/3000], train_loss: 2.2084, valid_loss: 1.4016

Epoch [1331/3000], train_loss: 2.1696, valid_loss: 1.5141

Epoch [1332/3000], train_loss: 2.0782, valid_loss: 1.7072

Epoch [1333/3000], train_loss: 2.7588, valid_loss: 1.7447

Epoch [1334/3000], train_loss: 2.3647, valid_loss: 1.4100

Epoch [1335/3000], train_loss: 2.2407, valid_loss: 1.4038

Epoch [1336/3000], train_loss: 2.0954, valid_loss: 1.6559

Epoch [1337/3000], train_loss: 2.0527, valid_loss: 1.5119

Epoch [1338/3000], train_loss: 2.1549, valid_loss: 1.6842

Epoch [1339/3000], train_loss: 2.0866, valid_loss: 1.4570

Epoch [1340/3000], train_loss: 2.0864, valid_loss: 1.3468

Epoch [1341/3000], train_loss: 2.3068, valid_loss: 1.6876

Epoch [1342/3000], train_loss: 2.0835, valid_loss: 1.5692

Epoch [1343/3000], train_loss: 2.5608, valid_loss: 1.7779

Epoch [1344/3000], train_loss: 2.3927, valid_loss: 1.3723

Epoch [1345/3000], train_loss: 2.0856, valid_loss: 1.6315

Epoch [1346/3000], train_loss: 2.5916, valid_loss: 1.4859

Epoch [1347/3000], train_loss: 2.0384, valid_loss: 1.3798

Epoch [1348/3000], train_loss: 2.4168, valid_loss: 1.4570

Epoch [1349/3000], train_loss: 2.3246, valid_loss: 1.7854

Epoch [1350/3000], train_loss: 2.1734, valid_loss: 1.7112

Epoch [1351/3000], train_loss: 2.2781, valid_loss: 1.5617

Epoch [1352/3000], train_loss: 2.1590, valid_loss: 1.4061

Epoch [1353/3000], train_loss: 2.3301, valid_loss: 1.4206

Epoch [1354/3000], train_loss: 2.2190, valid_loss: 1.5921

Epoch [1355/3000], train_loss: 2.2125, valid_loss: 1.4778

Epoch [1356/3000], train_loss: 2.2290, valid_loss: 1.3004

Epoch [1357/3000], train_loss: 2.6301, valid_loss: 1.8091

Epoch [1358/3000], train_loss: 2.2722, valid_loss: 1.5469

Epoch [1359/3000], train_loss: 2.2219, valid_loss: 1.3855

Epoch [1360/3000], train_loss: 2.3120, valid_loss: 1.3570

Epoch [1361/3000], train_loss: 2.3461, valid_loss: 1.3224

Epoch [1362/3000], train_loss: 2.1151, valid_loss: 1.5864

Epoch [1363/3000], train_loss: 2.3354, valid_loss: 1.4758

Epoch [1364/3000], train_loss: 2.1083, valid_loss: 1.4328

Epoch [1365/3000], train_loss: 2.3263, valid_loss: 1.6099

Epoch [1366/3000], train_loss: 2.3338, valid_loss: 1.5116

Epoch [1367/3000], train_loss: 2.3177, valid_loss: 1.5415

Epoch [1368/3000], train_loss: 2.2429, valid_loss: 1.5570

Epoch [1369/3000], train_loss: 2.3855, valid_loss: 1.6015

Epoch [1370/3000], train_loss: 2.1198, valid_loss: 1.6128

Epoch [1371/3000], train_loss: 2.2455, valid_loss: 1.5377

Epoch [1372/3000], train_loss: 2.2233, valid_loss: 1.8212

Epoch [1373/3000], train_loss: 2.1256, valid_loss: 1.7843

Epoch [1374/3000], train_loss: 2.3075, valid_loss: 1.4988

Epoch [1375/3000], train_loss: 2.4148, valid_loss: 1.7712

Epoch [1376/3000], train_loss: 2.5172, valid_loss: 1.4901

Epoch [1377/3000], train_loss: 2.3940, valid_loss: 1.4787

Epoch [1378/3000], train_loss: 2.2109, valid_loss: 1.3918

Epoch [1379/3000], train_loss: 2.1721, valid_loss: 1.5605

Epoch [1380/3000], train_loss: 2.3004, valid_loss: 1.6318

Epoch [1381/3000], train_loss: 2.7262, valid_loss: 1.2957

Epoch [1382/3000], train_loss: 2.2298, valid_loss: 1.2438
Save model at epoch 1382 with loss 1.2438

Epoch [1383/3000], train_loss: 1.9479, valid_loss: 1.8946

Epoch [1384/3000], train_loss: 2.0824, valid_loss: 1.7821

Epoch [1385/3000], train_loss: 2.0729, valid_loss: 1.5587

Epoch [1386/3000], train_loss: 2.1902, valid_loss: 1.4711

Epoch [1387/3000], train_loss: 2.3934, valid_loss: 1.5977

Epoch [1388/3000], train_loss: 2.1508, valid_loss: 1.4989

Epoch [1389/3000], train_loss: 2.1513, valid_loss: 1.5017

Epoch [1390/3000], train_loss: 2.1419, valid_loss: 1.3321

Epoch [1391/3000], train_loss: 2.1594, valid_loss: 1.3197

Epoch [1392/3000], train_loss: 2.4525, valid_loss: 1.2392
Save model at epoch 1392 with loss 1.2392

Epoch [1393/3000], train_loss: 2.0217, valid_loss: 1.5226

Epoch [1394/3000], train_loss: 2.2796, valid_loss: 1.5578

Epoch [1395/3000], train_loss: 2.1509, valid_loss: 1.3024

Epoch [1396/3000], train_loss: 2.2105, valid_loss: 1.7583

Epoch [1397/3000], train_loss: 2.3025, valid_loss: 1.4857

Epoch [1398/3000], train_loss: 2.3205, valid_loss: 1.6419

Epoch [1399/3000], train_loss: 2.2694, valid_loss: 1.5875

Epoch [1400/3000], train_loss: 1.9859, valid_loss: 1.6306

Epoch [1401/3000], train_loss: 2.0860, valid_loss: 1.5518

Epoch [1402/3000], train_loss: 2.3475, valid_loss: 2.0262

Epoch [1403/3000], train_loss: 2.2316, valid_loss: 1.8901

Epoch [1404/3000], train_loss: 2.3966, valid_loss: 1.4141

Epoch [1405/3000], train_loss: 2.0864, valid_loss: 1.5074

Epoch [1406/3000], train_loss: 2.0447, valid_loss: 1.4487

Epoch [1407/3000], train_loss: 2.1715, valid_loss: 1.4449

Epoch [1408/3000], train_loss: 2.3111, valid_loss: 1.7979

Epoch [1409/3000], train_loss: 2.0693, valid_loss: 1.4540

Epoch [1410/3000], train_loss: 2.2820, valid_loss: 1.5054

Epoch [1411/3000], train_loss: 2.2586, valid_loss: 1.4937

Epoch [1412/3000], train_loss: 2.2212, valid_loss: 1.5744

Epoch [1413/3000], train_loss: 2.1451, valid_loss: 1.5243

Epoch [1414/3000], train_loss: 1.9442, valid_loss: 1.4315

Epoch [1415/3000], train_loss: 2.2358, valid_loss: 1.3688

Epoch [1416/3000], train_loss: 2.1948, valid_loss: 1.6717

Epoch [1417/3000], train_loss: 2.1623, valid_loss: 1.4733

Epoch [1418/3000], train_loss: 2.4576, valid_loss: 1.3823

Epoch [1419/3000], train_loss: 2.4856, valid_loss: 1.4490

Epoch [1420/3000], train_loss: 2.5014, valid_loss: 1.4021

Epoch [1421/3000], train_loss: 2.1242, valid_loss: 1.3064

Epoch [1422/3000], train_loss: 2.0421, valid_loss: 1.3095

Epoch [1423/3000], train_loss: 2.2651, valid_loss: 1.6872

Epoch [1424/3000], train_loss: 2.2931, valid_loss: 1.5629

Epoch [1425/3000], train_loss: 2.4534, valid_loss: 1.7026

Epoch [1426/3000], train_loss: 2.1794, valid_loss: 1.3589

Epoch [1427/3000], train_loss: 2.3588, valid_loss: 2.0837

Epoch [1428/3000], train_loss: 2.1387, valid_loss: 1.7901

Epoch [1429/3000], train_loss: 2.3755, valid_loss: 1.3524

Epoch [1430/3000], train_loss: 2.3175, valid_loss: 1.2917

Epoch [1431/3000], train_loss: 2.1383, valid_loss: 1.7320

Epoch [1432/3000], train_loss: 2.5020, valid_loss: 1.3864

Epoch [1433/3000], train_loss: 2.1709, valid_loss: 1.4208

Epoch [1434/3000], train_loss: 2.2406, valid_loss: 1.6807

Epoch [1435/3000], train_loss: 1.9864, valid_loss: 1.5874

Epoch [1436/3000], train_loss: 2.1685, valid_loss: 1.4398

Epoch [1437/3000], train_loss: 2.0893, valid_loss: 1.3432

Epoch [1438/3000], train_loss: 2.0714, valid_loss: 1.5966

Epoch [1439/3000], train_loss: 2.2276, valid_loss: 1.5230

Epoch [1440/3000], train_loss: 2.0970, valid_loss: 1.4216

Epoch [1441/3000], train_loss: 2.1468, valid_loss: 1.4341

Epoch [1442/3000], train_loss: 2.1814, valid_loss: 1.3550

Epoch [1443/3000], train_loss: 2.2363, valid_loss: 1.4174

Epoch [1444/3000], train_loss: 2.1554, valid_loss: 1.2911

Epoch [1445/3000], train_loss: 2.1522, valid_loss: 1.5975

Epoch [1446/3000], train_loss: 2.3476, valid_loss: 1.7122

Epoch [1447/3000], train_loss: 2.0885, valid_loss: 1.6613

Epoch [1448/3000], train_loss: 1.9918, valid_loss: 1.4484

Epoch [1449/3000], train_loss: 1.9752, valid_loss: 1.3098

Epoch [1450/3000], train_loss: 2.0752, valid_loss: 1.6519

Epoch [1451/3000], train_loss: 2.3688, valid_loss: 1.5285

Epoch [1452/3000], train_loss: 1.9555, valid_loss: 1.4991

Epoch [1453/3000], train_loss: 2.3485, valid_loss: 1.5819

Epoch [1454/3000], train_loss: 2.1119, valid_loss: 1.8547

Epoch [1455/3000], train_loss: 2.1413, valid_loss: 1.6248

Epoch [1456/3000], train_loss: 2.1624, valid_loss: 1.5975

Epoch [1457/3000], train_loss: 2.1347, valid_loss: 1.5567

Epoch [1458/3000], train_loss: 2.1191, valid_loss: 1.4918

Epoch [1459/3000], train_loss: 2.1522, valid_loss: 1.4999

Epoch [1460/3000], train_loss: 2.0387, valid_loss: 1.3435

Epoch [1461/3000], train_loss: 2.2567, valid_loss: 1.4836

Epoch [1462/3000], train_loss: 2.3035, valid_loss: 1.4693

Epoch [1463/3000], train_loss: 2.0572, valid_loss: 1.2662

Epoch [1464/3000], train_loss: 2.2510, valid_loss: 1.4125

Epoch [1465/3000], train_loss: 2.0791, valid_loss: 1.6064

Epoch [1466/3000], train_loss: 2.2023, valid_loss: 1.5266

Epoch [1467/3000], train_loss: 2.1212, valid_loss: 1.2476

Epoch [1468/3000], train_loss: 2.4102, valid_loss: 1.4097

Epoch [1469/3000], train_loss: 2.1712, valid_loss: 1.4617

Epoch [1470/3000], train_loss: 2.2731, valid_loss: 1.4801

Epoch [1471/3000], train_loss: 1.8906, valid_loss: 1.6600

Epoch [1472/3000], train_loss: 2.0428, valid_loss: 1.6368

Epoch [1473/3000], train_loss: 2.2404, valid_loss: 1.2798

Epoch [1474/3000], train_loss: 2.3053, valid_loss: 1.4886

Epoch [1475/3000], train_loss: 2.2863, valid_loss: 1.6920

Epoch [1476/3000], train_loss: 1.9081, valid_loss: 1.2656

Epoch [1477/3000], train_loss: 2.0855, valid_loss: 1.4479

Epoch [1478/3000], train_loss: 2.7044, valid_loss: 1.4397

Epoch [1479/3000], train_loss: 2.1354, valid_loss: 1.4002

Epoch [1480/3000], train_loss: 2.3265, valid_loss: 1.2876

Epoch [1481/3000], train_loss: 2.0493, valid_loss: 1.6975

Epoch [1482/3000], train_loss: 2.1620, valid_loss: 1.5874

Epoch [1483/3000], train_loss: 1.9058, valid_loss: 1.4154

Epoch [1484/3000], train_loss: 2.3393, valid_loss: 1.6576

Epoch [1485/3000], train_loss: 2.0130, valid_loss: 1.5598

Epoch [1486/3000], train_loss: 2.2902, valid_loss: 1.7236

Epoch [1487/3000], train_loss: 2.7042, valid_loss: 1.4403

Epoch [1488/3000], train_loss: 2.2435, valid_loss: 1.2031
Save model at epoch 1488 with loss 1.2031

Epoch [1489/3000], train_loss: 2.2731, valid_loss: 1.4182

Epoch [1490/3000], train_loss: 2.1164, valid_loss: 1.4125

Epoch [1491/3000], train_loss: 2.1694, valid_loss: 1.6799

Epoch [1492/3000], train_loss: 1.9289, valid_loss: 1.3868

Epoch [1493/3000], train_loss: 2.2182, valid_loss: 1.6464

Epoch [1494/3000], train_loss: 2.1285, valid_loss: 1.2577

Epoch [1495/3000], train_loss: 2.2467, valid_loss: 1.4156

Epoch [1496/3000], train_loss: 2.0801, valid_loss: 1.2876

Epoch [1497/3000], train_loss: 2.3340, valid_loss: 1.4992

Epoch [1498/3000], train_loss: 2.5990, valid_loss: 1.5679

Epoch [1499/3000], train_loss: 2.6418, valid_loss: 1.3568

Epoch [1500/3000], train_loss: 2.2760, valid_loss: 1.4725

Epoch [1501/3000], train_loss: 1.9825, valid_loss: 1.4186

Epoch [1502/3000], train_loss: 2.4629, valid_loss: 1.7605

Epoch [1503/3000], train_loss: 2.2553, valid_loss: 1.5302

Epoch [1504/3000], train_loss: 2.1565, valid_loss: 1.4671

Epoch [1505/3000], train_loss: 2.2714, valid_loss: 1.4253

Epoch [1506/3000], train_loss: 1.8989, valid_loss: 1.5302

Epoch [1507/3000], train_loss: 2.0136, valid_loss: 1.2652

Epoch [1508/3000], train_loss: 1.9404, valid_loss: 1.4559

Epoch [1509/3000], train_loss: 2.1099, valid_loss: 1.3796

Epoch [1510/3000], train_loss: 2.0290, valid_loss: 1.4670

Epoch [1511/3000], train_loss: 2.3950, valid_loss: 1.5949

Epoch [1512/3000], train_loss: 2.0754, valid_loss: 1.5655

Epoch [1513/3000], train_loss: 2.3038, valid_loss: 1.3267

Epoch [1514/3000], train_loss: 2.2901, valid_loss: 1.4130

Epoch [1515/3000], train_loss: 2.0196, valid_loss: 1.6026

Epoch [1516/3000], train_loss: 2.0668, valid_loss: 1.4114

Epoch [1517/3000], train_loss: 1.9866, valid_loss: 1.2348

Epoch [1518/3000], train_loss: 2.1927, valid_loss: 1.7404

Epoch [1519/3000], train_loss: 2.1485, valid_loss: 1.4746

Epoch [1520/3000], train_loss: 2.2460, valid_loss: 1.6937

Epoch [1521/3000], train_loss: 1.9665, valid_loss: 1.3327

Epoch [1522/3000], train_loss: 2.1820, valid_loss: 1.4458

Epoch [1523/3000], train_loss: 2.5660, valid_loss: 1.5101

Epoch [1524/3000], train_loss: 2.1612, valid_loss: 1.4321

Epoch [1525/3000], train_loss: 2.1641, valid_loss: 1.5587

Epoch [1526/3000], train_loss: 2.1150, valid_loss: 1.4766

Epoch [1527/3000], train_loss: 2.1577, valid_loss: 1.2542

Epoch [1528/3000], train_loss: 2.5034, valid_loss: 1.4526

Epoch [1529/3000], train_loss: 1.9761, valid_loss: 1.5238

Epoch [1530/3000], train_loss: 2.1411, valid_loss: 1.4715

Epoch [1531/3000], train_loss: 2.4054, valid_loss: 1.4607

Epoch [1532/3000], train_loss: 2.1235, valid_loss: 1.6191

Epoch [1533/3000], train_loss: 2.0533, valid_loss: 1.2914

Epoch [1534/3000], train_loss: 2.0917, valid_loss: 1.4227

Epoch [1535/3000], train_loss: 2.0624, valid_loss: 1.3578

Epoch [1536/3000], train_loss: 2.1227, valid_loss: 1.3378

Epoch [1537/3000], train_loss: 2.1650, valid_loss: 1.5729

Epoch [1538/3000], train_loss: 2.0570, valid_loss: 1.4520

Epoch [1539/3000], train_loss: 2.3119, valid_loss: 1.6188

Epoch [1540/3000], train_loss: 2.2437, valid_loss: 1.4040

Epoch [1541/3000], train_loss: 2.2056, valid_loss: 1.5992

Epoch [1542/3000], train_loss: 2.2396, valid_loss: 1.3177

Epoch [1543/3000], train_loss: 2.4646, valid_loss: 1.2751

Epoch [1544/3000], train_loss: 1.9605, valid_loss: 1.3592

Epoch [1545/3000], train_loss: 2.2295, valid_loss: 1.5271

Epoch [1546/3000], train_loss: 2.3017, valid_loss: 1.4063

Epoch [1547/3000], train_loss: 2.1920, valid_loss: 1.2779

Epoch [1548/3000], train_loss: 2.2906, valid_loss: 1.2941

Epoch [1549/3000], train_loss: 2.0630, valid_loss: 1.2282

Epoch [1550/3000], train_loss: 2.0416, valid_loss: 1.4526

Epoch [1551/3000], train_loss: 2.0690, valid_loss: 1.4500

Epoch [1552/3000], train_loss: 2.1457, valid_loss: 1.6286

Epoch [1553/3000], train_loss: 2.1323, valid_loss: 1.4276

Epoch [1554/3000], train_loss: 2.0674, valid_loss: 1.5951

Epoch [1555/3000], train_loss: 2.1775, valid_loss: 1.3425

Epoch [1556/3000], train_loss: 2.2629, valid_loss: 1.2634

Epoch [1557/3000], train_loss: 1.9113, valid_loss: 1.8535

Epoch [1558/3000], train_loss: 2.5419, valid_loss: 1.3341

Epoch [1559/3000], train_loss: 2.1005, valid_loss: 1.4793

Epoch [1560/3000], train_loss: 2.2841, valid_loss: 1.4292

Epoch [1561/3000], train_loss: 1.9731, valid_loss: 1.3897

Epoch [1562/3000], train_loss: 2.0547, valid_loss: 1.3576

Epoch [1563/3000], train_loss: 2.1451, valid_loss: 1.8255

Epoch [1564/3000], train_loss: 2.0494, valid_loss: 1.5851

Epoch [1565/3000], train_loss: 2.3108, valid_loss: 1.3679

Epoch [1566/3000], train_loss: 2.0165, valid_loss: 1.4100

Epoch [1567/3000], train_loss: 1.9958, valid_loss: 1.4968

Epoch [1568/3000], train_loss: 2.4186, valid_loss: 1.4964

Epoch [1569/3000], train_loss: 2.0365, valid_loss: 1.2867

Epoch [1570/3000], train_loss: 2.2372, valid_loss: 1.6731

Epoch [1571/3000], train_loss: 2.2302, valid_loss: 1.2698

Epoch [1572/3000], train_loss: 2.0353, valid_loss: 1.3289

Epoch [1573/3000], train_loss: 1.9575, valid_loss: 1.5878

Epoch [1574/3000], train_loss: 2.0062, valid_loss: 1.2634

Epoch [1575/3000], train_loss: 2.1277, valid_loss: 1.4950

Epoch [1576/3000], train_loss: 2.2616, valid_loss: 1.6469

Epoch [1577/3000], train_loss: 2.1189, valid_loss: 1.4163

Epoch [1578/3000], train_loss: 2.0463, valid_loss: 1.4574

Epoch [1579/3000], train_loss: 2.2688, valid_loss: 1.3887

Epoch [1580/3000], train_loss: 1.9995, valid_loss: 1.4289

Epoch [1581/3000], train_loss: 2.0741, valid_loss: 1.4904

Epoch [1582/3000], train_loss: 2.1246, valid_loss: 1.4984

Epoch [1583/3000], train_loss: 2.0551, valid_loss: 1.3135

Epoch [1584/3000], train_loss: 2.0132, valid_loss: 1.4998

Epoch [1585/3000], train_loss: 2.1340, valid_loss: 1.6612

Epoch [1586/3000], train_loss: 1.9916, valid_loss: 1.2968

Epoch [1587/3000], train_loss: 2.2233, valid_loss: 1.6994

Epoch [1588/3000], train_loss: 2.2217, valid_loss: 1.4816

Epoch [1589/3000], train_loss: 2.0687, valid_loss: 1.4504

Epoch [1590/3000], train_loss: 2.1019, valid_loss: 1.4345

Epoch [1591/3000], train_loss: 1.9329, valid_loss: 1.4195

Epoch [1592/3000], train_loss: 2.3147, valid_loss: 1.4202

Epoch [1593/3000], train_loss: 2.2137, valid_loss: 1.3605

Epoch [1594/3000], train_loss: 2.0403, valid_loss: 1.5037

Epoch [1595/3000], train_loss: 2.0505, valid_loss: 1.3974

Epoch [1596/3000], train_loss: 2.0860, valid_loss: 1.5406

Epoch [1597/3000], train_loss: 2.2125, valid_loss: 1.4339

Epoch [1598/3000], train_loss: 2.0217, valid_loss: 1.4452

Epoch [1599/3000], train_loss: 2.0712, valid_loss: 1.4329

Epoch [1600/3000], train_loss: 2.0828, valid_loss: 1.2771

Epoch [1601/3000], train_loss: 1.9438, valid_loss: 1.3187

Epoch [1602/3000], train_loss: 2.2414, valid_loss: 1.2892

Epoch [1603/3000], train_loss: 2.2885, valid_loss: 1.3890

Epoch [1604/3000], train_loss: 2.0913, valid_loss: 1.3670

Epoch [1605/3000], train_loss: 2.3954, valid_loss: 1.2892

Epoch [1606/3000], train_loss: 2.0763, valid_loss: 1.4847

Epoch [1607/3000], train_loss: 2.2194, valid_loss: 1.2871

Epoch [1608/3000], train_loss: 2.2465, valid_loss: 1.3881

Epoch [1609/3000], train_loss: 1.9860, valid_loss: 2.0339

Epoch [1610/3000], train_loss: 2.1823, valid_loss: 1.5332

Epoch [1611/3000], train_loss: 2.1004, valid_loss: 1.3845

Epoch [1612/3000], train_loss: 2.0021, valid_loss: 1.3625

Epoch [1613/3000], train_loss: 2.0929, valid_loss: 1.3916

Epoch [1614/3000], train_loss: 2.2283, valid_loss: 1.7005

Epoch [1615/3000], train_loss: 2.0989, valid_loss: 1.8611

Epoch [1616/3000], train_loss: 2.1769, valid_loss: 1.4872

Epoch [1617/3000], train_loss: 2.1718, valid_loss: 1.3863

Epoch [1618/3000], train_loss: 1.8899, valid_loss: 1.4549

Epoch [1619/3000], train_loss: 2.2624, valid_loss: 1.4730

Epoch [1620/3000], train_loss: 2.0550, valid_loss: 1.5734

Epoch [1621/3000], train_loss: 1.8399, valid_loss: 1.3850

Epoch [1622/3000], train_loss: 2.2425, valid_loss: 1.1875
Save model at epoch 1622 with loss 1.1875

Epoch [1623/3000], train_loss: 2.0239, valid_loss: 1.7499

Epoch [1624/3000], train_loss: 2.1773, valid_loss: 1.6061

Epoch [1625/3000], train_loss: 1.9165, valid_loss: 1.4620

Epoch [1626/3000], train_loss: 1.9625, valid_loss: 1.5140

Epoch [1627/3000], train_loss: 2.8600, valid_loss: 1.5148

Epoch [1628/3000], train_loss: 1.9696, valid_loss: 1.6281

Epoch [1629/3000], train_loss: 2.0281, valid_loss: 1.3595

Epoch [1630/3000], train_loss: 2.2712, valid_loss: 1.5059

Epoch [1631/3000], train_loss: 1.9568, valid_loss: 1.4831

Epoch [1632/3000], train_loss: 2.0687, valid_loss: 1.6464

Epoch [1633/3000], train_loss: 2.0201, valid_loss: 1.3335

Epoch [1634/3000], train_loss: 2.3003, valid_loss: 1.3665

Epoch [1635/3000], train_loss: 2.5545, valid_loss: 1.3759

Epoch [1636/3000], train_loss: 1.9210, valid_loss: 1.4070

Epoch [1637/3000], train_loss: 2.2139, valid_loss: 1.2492

Epoch [1638/3000], train_loss: 1.8087, valid_loss: 1.3721

Epoch [1639/3000], train_loss: 2.0020, valid_loss: 1.4396

Epoch [1640/3000], train_loss: 2.2169, valid_loss: 1.2167

Epoch [1641/3000], train_loss: 2.4434, valid_loss: 1.2970

Epoch [1642/3000], train_loss: 1.9682, valid_loss: 1.4342

Epoch [1643/3000], train_loss: 2.0073, valid_loss: 1.4482

Epoch [1644/3000], train_loss: 2.1545, valid_loss: 1.6253

Epoch [1645/3000], train_loss: 2.0754, valid_loss: 1.4072

Epoch [1646/3000], train_loss: 2.3329, valid_loss: 1.3054

Epoch [1647/3000], train_loss: 2.2887, valid_loss: 1.2967

Epoch [1648/3000], train_loss: 2.0136, valid_loss: 1.5230

Epoch [1649/3000], train_loss: 2.2865, valid_loss: 1.2637

Epoch [1650/3000], train_loss: 2.0810, valid_loss: 1.5168

Epoch [1651/3000], train_loss: 2.5683, valid_loss: 1.3535

Epoch [1652/3000], train_loss: 2.3221, valid_loss: 1.5908

Epoch [1653/3000], train_loss: 2.0992, valid_loss: 1.7096

Epoch [1654/3000], train_loss: 1.8657, valid_loss: 1.4195

Epoch [1655/3000], train_loss: 2.0773, valid_loss: 1.3941

Epoch [1656/3000], train_loss: 2.0600, valid_loss: 1.2687

Epoch [1657/3000], train_loss: 1.9238, valid_loss: 1.2591

Epoch [1658/3000], train_loss: 1.9233, valid_loss: 1.4254

Epoch [1659/3000], train_loss: 1.8084, valid_loss: 1.6238

Epoch [1660/3000], train_loss: 2.0645, valid_loss: 1.2733

Epoch [1661/3000], train_loss: 2.1070, valid_loss: 1.4081

Epoch [1662/3000], train_loss: 2.1634, valid_loss: 1.3427

Epoch [1663/3000], train_loss: 2.0225, valid_loss: 1.2981

Epoch [1664/3000], train_loss: 2.1081, valid_loss: 1.4476

Epoch [1665/3000], train_loss: 2.1005, valid_loss: 1.2429

Epoch [1666/3000], train_loss: 2.0412, valid_loss: 1.4358

Epoch [1667/3000], train_loss: 1.9718, valid_loss: 1.6197

Epoch [1668/3000], train_loss: 2.1697, valid_loss: 1.4008

Epoch [1669/3000], train_loss: 2.1293, valid_loss: 1.2606

Epoch [1670/3000], train_loss: 2.1062, valid_loss: 1.5256

Epoch [1671/3000], train_loss: 1.8885, valid_loss: 1.4500

Epoch [1672/3000], train_loss: 1.9665, valid_loss: 1.5319

Epoch [1673/3000], train_loss: 2.3758, valid_loss: 1.3534

Epoch [1674/3000], train_loss: 2.1975, valid_loss: 1.4245

Epoch [1675/3000], train_loss: 2.3204, valid_loss: 1.6766

Epoch [1676/3000], train_loss: 1.9583, valid_loss: 1.3592

Epoch [1677/3000], train_loss: 2.0491, valid_loss: 1.2791

Epoch [1678/3000], train_loss: 2.0208, valid_loss: 1.3927

Epoch [1679/3000], train_loss: 2.0923, valid_loss: 1.4163

Epoch [1680/3000], train_loss: 1.9245, valid_loss: 1.5151

Epoch [1681/3000], train_loss: 1.9379, valid_loss: 1.3955

Epoch [1682/3000], train_loss: 2.4770, valid_loss: 1.1670
Save model at epoch 1682 with loss 1.1670

Epoch [1683/3000], train_loss: 2.0387, valid_loss: 1.2771

Epoch [1684/3000], train_loss: 2.1888, valid_loss: 1.2923

Epoch [1685/3000], train_loss: 2.1234, valid_loss: 1.2982

Epoch [1686/3000], train_loss: 2.0500, valid_loss: 1.5583

Epoch [1687/3000], train_loss: 2.0942, valid_loss: 1.5177

Epoch [1688/3000], train_loss: 2.4385, valid_loss: 1.8164

Epoch [1689/3000], train_loss: 2.1100, valid_loss: 1.4468

Epoch [1690/3000], train_loss: 2.3205, valid_loss: 1.6234

Epoch [1691/3000], train_loss: 2.0102, valid_loss: 1.4917

Epoch [1692/3000], train_loss: 1.9779, valid_loss: 1.6877

Epoch [1693/3000], train_loss: 1.9391, valid_loss: 1.3591

Epoch [1694/3000], train_loss: 1.9819, valid_loss: 1.4330

Epoch [1695/3000], train_loss: 2.1000, valid_loss: 1.4816

Epoch [1696/3000], train_loss: 2.0323, valid_loss: 1.3139

Epoch [1697/3000], train_loss: 1.9955, valid_loss: 1.3171

Epoch [1698/3000], train_loss: 2.0648, valid_loss: 1.2316

Epoch [1699/3000], train_loss: 2.1135, valid_loss: 1.3855

Epoch [1700/3000], train_loss: 2.4607, valid_loss: 1.3521

Epoch [1701/3000], train_loss: 2.0304, valid_loss: 1.3298

Epoch [1702/3000], train_loss: 1.9564, valid_loss: 1.2324

Epoch [1703/3000], train_loss: 2.0552, valid_loss: 1.8634

Epoch [1704/3000], train_loss: 1.9339, valid_loss: 1.1727

Epoch [1705/3000], train_loss: 2.0044, valid_loss: 1.3423

Epoch [1706/3000], train_loss: 2.0741, valid_loss: 1.3823

Epoch [1707/3000], train_loss: 2.1934, valid_loss: 1.3293

Epoch [1708/3000], train_loss: 2.3414, valid_loss: 1.4624

Epoch [1709/3000], train_loss: 2.1782, valid_loss: 1.2640

Epoch [1710/3000], train_loss: 2.0171, valid_loss: 1.6361

Epoch [1711/3000], train_loss: 2.1494, valid_loss: 1.2435

Epoch [1712/3000], train_loss: 1.8174, valid_loss: 1.3564

Epoch [1713/3000], train_loss: 1.9749, valid_loss: 1.5077

Epoch [1714/3000], train_loss: 2.0029, valid_loss: 1.2307

Epoch [1715/3000], train_loss: 1.9111, valid_loss: 1.3831

Epoch [1716/3000], train_loss: 1.9956, valid_loss: 1.4483

Epoch [1717/3000], train_loss: 2.3781, valid_loss: 1.4356

Epoch [1718/3000], train_loss: 2.1448, valid_loss: 1.3550

Epoch [1719/3000], train_loss: 2.0002, valid_loss: 1.3107

Epoch [1720/3000], train_loss: 1.9926, valid_loss: 1.5294

Epoch [1721/3000], train_loss: 2.2693, valid_loss: 1.6294

Epoch [1722/3000], train_loss: 2.0026, valid_loss: 1.5954

Epoch [1723/3000], train_loss: 2.3647, valid_loss: 1.2196

Epoch [1724/3000], train_loss: 2.1779, valid_loss: 1.3035

Epoch [1725/3000], train_loss: 1.9508, valid_loss: 1.3990

Epoch [1726/3000], train_loss: 1.9425, valid_loss: 1.3215

Epoch [1727/3000], train_loss: 2.0481, valid_loss: 1.5268

Epoch [1728/3000], train_loss: 2.0135, valid_loss: 1.5960

Epoch [1729/3000], train_loss: 1.8711, valid_loss: 1.4153

Epoch [1730/3000], train_loss: 2.3776, valid_loss: 1.5692

Epoch [1731/3000], train_loss: 1.8973, valid_loss: 1.3455

Epoch [1732/3000], train_loss: 1.9886, valid_loss: 1.4406

Epoch [1733/3000], train_loss: 2.0333, valid_loss: 1.3581

Epoch [1734/3000], train_loss: 2.1739, valid_loss: 1.3915

Epoch [1735/3000], train_loss: 2.2612, valid_loss: 1.2538

Epoch [1736/3000], train_loss: 2.0372, valid_loss: 1.3765

Epoch [1737/3000], train_loss: 1.9041, valid_loss: 1.6274

Epoch [1738/3000], train_loss: 1.9829, valid_loss: 1.3020

Epoch [1739/3000], train_loss: 2.1231, valid_loss: 1.4259

Epoch [1740/3000], train_loss: 1.9281, valid_loss: 1.5150

Epoch [1741/3000], train_loss: 2.0955, valid_loss: 1.3669

Epoch [1742/3000], train_loss: 2.0226, valid_loss: 1.4096

Epoch [1743/3000], train_loss: 2.5642, valid_loss: 1.1874

Epoch [1744/3000], train_loss: 1.8604, valid_loss: 1.3646

Epoch [1745/3000], train_loss: 1.9514, valid_loss: 1.4188

Epoch [1746/3000], train_loss: 2.4716, valid_loss: 1.5949

Epoch [1747/3000], train_loss: 2.1690, valid_loss: 1.2893

Epoch [1748/3000], train_loss: 2.1638, valid_loss: 1.6108

Epoch [1749/3000], train_loss: 1.9079, valid_loss: 1.4260

Epoch [1750/3000], train_loss: 2.0059, valid_loss: 1.5356

Epoch [1751/3000], train_loss: 2.0250, valid_loss: 1.6534

Epoch [1752/3000], train_loss: 1.9478, valid_loss: 1.4745

Epoch [1753/3000], train_loss: 1.9594, valid_loss: 1.3229

Epoch [1754/3000], train_loss: 2.2082, valid_loss: 1.4408

Epoch [1755/3000], train_loss: 1.8715, valid_loss: 1.1907

Epoch [1756/3000], train_loss: 2.0906, valid_loss: 1.3256

Epoch [1757/3000], train_loss: 2.1000, valid_loss: 1.6210

Epoch [1758/3000], train_loss: 2.1682, valid_loss: 1.6344

Epoch [1759/3000], train_loss: 2.0195, valid_loss: 1.4758

Epoch [1760/3000], train_loss: 1.9595, valid_loss: 1.2272

Epoch [1761/3000], train_loss: 1.9715, valid_loss: 1.4215

Epoch [1762/3000], train_loss: 1.7318, valid_loss: 1.3062

Epoch [1763/3000], train_loss: 2.0433, valid_loss: 1.3497

Epoch [1764/3000], train_loss: 1.8498, valid_loss: 1.5078

Epoch [1765/3000], train_loss: 1.9085, valid_loss: 1.3912

Epoch [1766/3000], train_loss: 1.9532, valid_loss: 1.2153

Epoch [1767/3000], train_loss: 2.0514, valid_loss: 1.4989

Epoch [1768/3000], train_loss: 2.3929, valid_loss: 1.6159

Epoch [1769/3000], train_loss: 1.9962, valid_loss: 1.2879

Epoch [1770/3000], train_loss: 1.9095, valid_loss: 1.7989

Epoch [1771/3000], train_loss: 2.3999, valid_loss: 1.3950

Epoch [1772/3000], train_loss: 2.0828, valid_loss: 1.4316

Epoch [1773/3000], train_loss: 1.9323, valid_loss: 1.5172

Epoch [1774/3000], train_loss: 1.9145, valid_loss: 1.2891

Epoch [1775/3000], train_loss: 2.0329, valid_loss: 1.5143

Epoch [1776/3000], train_loss: 1.8641, valid_loss: 1.2659

Epoch [1777/3000], train_loss: 1.9652, valid_loss: 1.2981

Epoch [1778/3000], train_loss: 1.9169, valid_loss: 1.1819

Epoch [1779/3000], train_loss: 1.9443, valid_loss: 1.4910

Epoch [1780/3000], train_loss: 1.9706, valid_loss: 1.5509

Epoch [1781/3000], train_loss: 2.3174, valid_loss: 1.2476

Epoch [1782/3000], train_loss: 2.4617, valid_loss: 1.3079

Epoch [1783/3000], train_loss: 2.0103, valid_loss: 1.7455

Epoch [1784/3000], train_loss: 1.8271, valid_loss: 1.3163

Epoch [1785/3000], train_loss: 1.9046, valid_loss: 1.5216

Epoch [1786/3000], train_loss: 1.9858, valid_loss: 1.3165

Epoch [1787/3000], train_loss: 2.1038, valid_loss: 1.5878

Epoch [1788/3000], train_loss: 1.9113, valid_loss: 1.4918

Epoch [1789/3000], train_loss: 1.8645, valid_loss: 1.3067

Epoch [1790/3000], train_loss: 2.1233, valid_loss: 1.4154

Epoch [1791/3000], train_loss: 1.8769, valid_loss: 1.3250

Epoch [1792/3000], train_loss: 1.9214, valid_loss: 1.6615

Epoch [1793/3000], train_loss: 2.2241, valid_loss: 1.5539

Epoch [1794/3000], train_loss: 2.0100, valid_loss: 1.4999

Epoch [1795/3000], train_loss: 2.1456, valid_loss: 1.3446

Epoch [1796/3000], train_loss: 1.8098, valid_loss: 1.5337

Epoch [1797/3000], train_loss: 2.1239, valid_loss: 1.3814

Epoch [1798/3000], train_loss: 1.8536, valid_loss: 1.3590

Epoch [1799/3000], train_loss: 2.2123, valid_loss: 1.4045

Epoch [1800/3000], train_loss: 1.8864, valid_loss: 1.5622

Epoch [1801/3000], train_loss: 2.3248, valid_loss: 1.3817

Epoch [1802/3000], train_loss: 1.8921, valid_loss: 1.3537

Epoch [1803/3000], train_loss: 2.0153, valid_loss: 1.4726

Epoch [1804/3000], train_loss: 1.7867, valid_loss: 1.2456

Epoch [1805/3000], train_loss: 2.1285, valid_loss: 1.3029

Epoch [1806/3000], train_loss: 1.8948, valid_loss: 1.1030
Save model at epoch 1806 with loss 1.1030

Epoch [1807/3000], train_loss: 2.2011, valid_loss: 1.4123

Epoch [1808/3000], train_loss: 1.8751, valid_loss: 1.2087

Epoch [1809/3000], train_loss: 2.0688, valid_loss: 1.4839

Epoch [1810/3000], train_loss: 2.3287, valid_loss: 1.9769

Epoch [1811/3000], train_loss: 1.8893, valid_loss: 1.3524

Epoch [1812/3000], train_loss: 1.8555, valid_loss: 1.3616

Epoch [1813/3000], train_loss: 2.0267, valid_loss: 1.4556

Epoch [1814/3000], train_loss: 1.9670, valid_loss: 1.2144

Epoch [1815/3000], train_loss: 1.7459, valid_loss: 1.4501

Epoch [1816/3000], train_loss: 1.9698, valid_loss: 1.3726

Epoch [1817/3000], train_loss: 1.8997, valid_loss: 1.2758

Epoch [1818/3000], train_loss: 1.8903, valid_loss: 1.5060

Epoch [1819/3000], train_loss: 1.9484, valid_loss: 1.8242

Epoch [1820/3000], train_loss: 2.1594, valid_loss: 1.3412

Epoch [1821/3000], train_loss: 1.9757, valid_loss: 1.3245

Epoch [1822/3000], train_loss: 1.8220, valid_loss: 1.4835

Epoch [1823/3000], train_loss: 1.9394, valid_loss: 1.6634

Epoch [1824/3000], train_loss: 1.9593, valid_loss: 1.2348

Epoch [1825/3000], train_loss: 2.0050, valid_loss: 1.3541

Epoch [1826/3000], train_loss: 2.0345, valid_loss: 1.4021

Epoch [1827/3000], train_loss: 2.1019, valid_loss: 1.3787

Epoch [1828/3000], train_loss: 2.0272, valid_loss: 1.3115

Epoch [1829/3000], train_loss: 2.0884, valid_loss: 1.4152

Epoch [1830/3000], train_loss: 1.9608, valid_loss: 1.2190

Epoch [1831/3000], train_loss: 1.9704, valid_loss: 1.4653

Epoch [1832/3000], train_loss: 2.3265, valid_loss: 1.6019

Epoch [1833/3000], train_loss: 2.0061, valid_loss: 1.5211

Epoch [1834/3000], train_loss: 2.0503, valid_loss: 1.3487

Epoch [1835/3000], train_loss: 2.0609, valid_loss: 1.4709

Epoch [1836/3000], train_loss: 2.1107, valid_loss: 1.5136

Epoch [1837/3000], train_loss: 1.7970, valid_loss: 1.3482

Epoch [1838/3000], train_loss: 2.0713, valid_loss: 1.2314

Epoch [1839/3000], train_loss: 2.0985, valid_loss: 1.5766

Epoch [1840/3000], train_loss: 2.0106, valid_loss: 1.2201

Epoch [1841/3000], train_loss: 1.8974, valid_loss: 1.3767

Epoch [1842/3000], train_loss: 1.8174, valid_loss: 1.2453

Epoch [1843/3000], train_loss: 1.8718, valid_loss: 1.3203

Epoch [1844/3000], train_loss: 1.8530, valid_loss: 1.4646

Epoch [1845/3000], train_loss: 2.2359, valid_loss: 1.6064

Epoch [1846/3000], train_loss: 2.2247, valid_loss: 1.4306

Epoch [1847/3000], train_loss: 2.0988, valid_loss: 1.4130

Epoch [1848/3000], train_loss: 1.8434, valid_loss: 1.7905

Epoch [1849/3000], train_loss: 2.0238, valid_loss: 1.3292

Epoch [1850/3000], train_loss: 1.8886, valid_loss: 1.4077

Epoch [1851/3000], train_loss: 1.9804, valid_loss: 1.6544

Epoch [1852/3000], train_loss: 1.7666, valid_loss: 1.4107

Epoch [1853/3000], train_loss: 1.9838, valid_loss: 1.3674

Epoch [1854/3000], train_loss: 1.9320, valid_loss: 1.2334

Epoch [1855/3000], train_loss: 1.8623, valid_loss: 1.4812

Epoch [1856/3000], train_loss: 2.2082, valid_loss: 1.4118

Epoch [1857/3000], train_loss: 1.9099, valid_loss: 1.4729

Epoch [1858/3000], train_loss: 2.1288, valid_loss: 1.3752

Epoch [1859/3000], train_loss: 2.1647, valid_loss: 1.5194

Epoch [1860/3000], train_loss: 1.9179, valid_loss: 1.3776

Epoch [1861/3000], train_loss: 2.0231, valid_loss: 1.3023

Epoch [1862/3000], train_loss: 2.1628, valid_loss: 1.2850

Epoch [1863/3000], train_loss: 2.0381, valid_loss: 1.4086

Epoch [1864/3000], train_loss: 1.9546, valid_loss: 1.3514

Epoch [1865/3000], train_loss: 1.9207, valid_loss: 1.3670

Epoch [1866/3000], train_loss: 1.9572, valid_loss: 1.5150

Epoch [1867/3000], train_loss: 1.9929, valid_loss: 1.4343

Epoch [1868/3000], train_loss: 1.7283, valid_loss: 1.6300

Epoch [1869/3000], train_loss: 1.9103, valid_loss: 1.2511

Epoch [1870/3000], train_loss: 2.1212, valid_loss: 1.2686

Epoch [1871/3000], train_loss: 1.8205, valid_loss: 1.3374

Epoch [1872/3000], train_loss: 2.1156, valid_loss: 1.3907

Epoch [1873/3000], train_loss: 2.0316, valid_loss: 1.4672

Epoch [1874/3000], train_loss: 2.0969, valid_loss: 1.5583

Epoch [1875/3000], train_loss: 1.9326, valid_loss: 1.4784

Epoch [1876/3000], train_loss: 2.0400, valid_loss: 1.2410

Epoch [1877/3000], train_loss: 1.9235, valid_loss: 1.3960

Epoch [1878/3000], train_loss: 1.9713, valid_loss: 1.5049

Epoch [1879/3000], train_loss: 2.1569, valid_loss: 1.4319

Epoch [1880/3000], train_loss: 1.9471, valid_loss: 1.2681

Epoch [1881/3000], train_loss: 1.9918, valid_loss: 1.7356

Epoch [1882/3000], train_loss: 1.9941, valid_loss: 1.5176

Epoch [1883/3000], train_loss: 1.8652, valid_loss: 1.2862

Epoch [1884/3000], train_loss: 1.9622, valid_loss: 1.2830

Epoch [1885/3000], train_loss: 1.8381, valid_loss: 1.3611

Epoch [1886/3000], train_loss: 2.0865, valid_loss: 1.5184

Epoch [1887/3000], train_loss: 2.0041, valid_loss: 1.3152

Epoch [1888/3000], train_loss: 2.1373, valid_loss: 1.2371

Epoch [1889/3000], train_loss: 1.8587, valid_loss: 1.6544

Epoch [1890/3000], train_loss: 2.1238, valid_loss: 1.2088

Epoch [1891/3000], train_loss: 1.9580, valid_loss: 1.2974

Epoch [1892/3000], train_loss: 1.8299, valid_loss: 1.3236

Epoch [1893/3000], train_loss: 1.7651, valid_loss: 1.3256

Epoch [1894/3000], train_loss: 1.9139, valid_loss: 1.3150

Epoch [1895/3000], train_loss: 2.0341, valid_loss: 1.4600

Epoch [1896/3000], train_loss: 1.8838, valid_loss: 1.5019

Epoch [1897/3000], train_loss: 2.0243, valid_loss: 1.5477

Epoch [1898/3000], train_loss: 1.8578, valid_loss: 1.2506

Epoch [1899/3000], train_loss: 1.9250, valid_loss: 1.2673

Epoch [1900/3000], train_loss: 2.0101, valid_loss: 1.4983

Epoch [1901/3000], train_loss: 1.8522, valid_loss: 1.3870

Epoch [1902/3000], train_loss: 1.9740, valid_loss: 1.3329

Epoch [1903/3000], train_loss: 1.9212, valid_loss: 1.1894

Epoch [1904/3000], train_loss: 1.8504, valid_loss: 1.2743

Epoch [1905/3000], train_loss: 1.8225, valid_loss: 1.2431

Epoch [1906/3000], train_loss: 2.0256, valid_loss: 1.3446

Epoch [1907/3000], train_loss: 1.9548, valid_loss: 1.5116

Epoch [1908/3000], train_loss: 1.8022, valid_loss: 1.4456

Epoch [1909/3000], train_loss: 2.2908, valid_loss: 1.2415

Epoch [1910/3000], train_loss: 1.9808, valid_loss: 1.2153

Epoch [1911/3000], train_loss: 1.8734, valid_loss: 1.2359

Epoch [1912/3000], train_loss: 1.9530, valid_loss: 1.3025

Epoch [1913/3000], train_loss: 2.0269, valid_loss: 1.3498

Epoch [1914/3000], train_loss: 1.9055, valid_loss: 1.3176

Epoch [1915/3000], train_loss: 1.8814, valid_loss: 1.1767

Epoch [1916/3000], train_loss: 1.9764, valid_loss: 1.1612

Epoch [1917/3000], train_loss: 1.7347, valid_loss: 1.3898

Epoch [1918/3000], train_loss: 1.9220, valid_loss: 1.3254

Epoch [1919/3000], train_loss: 1.8239, valid_loss: 1.5829

Epoch [1920/3000], train_loss: 2.0905, valid_loss: 1.4395

Epoch [1921/3000], train_loss: 1.8527, valid_loss: 1.4526

Epoch [1922/3000], train_loss: 1.9358, valid_loss: 1.2432

Epoch [1923/3000], train_loss: 1.9630, valid_loss: 1.3599

Epoch [1924/3000], train_loss: 1.9715, valid_loss: 1.2284

Epoch [1925/3000], train_loss: 1.8127, valid_loss: 1.4880

Epoch [1926/3000], train_loss: 1.8184, valid_loss: 1.5486

Epoch [1927/3000], train_loss: 1.8470, valid_loss: 1.2976

Epoch [1928/3000], train_loss: 2.1148, valid_loss: 1.3181

Epoch [1929/3000], train_loss: 1.8591, valid_loss: 1.3043

Epoch [1930/3000], train_loss: 2.0740, valid_loss: 1.3379

Epoch [1931/3000], train_loss: 1.8883, valid_loss: 1.4786

Epoch [1932/3000], train_loss: 2.0650, valid_loss: 1.2286

Epoch [1933/3000], train_loss: 1.8787, valid_loss: 1.5110

Epoch [1934/3000], train_loss: 2.0549, valid_loss: 1.6259

Epoch [1935/3000], train_loss: 2.1462, valid_loss: 1.4208

Epoch [1936/3000], train_loss: 1.8756, valid_loss: 1.2765

Epoch [1937/3000], train_loss: 1.9073, valid_loss: 1.4266

Epoch [1938/3000], train_loss: 2.1748, valid_loss: 1.7551

Epoch [1939/3000], train_loss: 2.1615, valid_loss: 1.6342

Epoch [1940/3000], train_loss: 1.9669, valid_loss: 1.7747

Epoch [1941/3000], train_loss: 1.9395, valid_loss: 1.5482

Epoch [1942/3000], train_loss: 1.8339, valid_loss: 1.1891

Epoch [1943/3000], train_loss: 1.8482, valid_loss: 1.4986

Epoch [1944/3000], train_loss: 1.9691, valid_loss: 1.6376

Epoch [1945/3000], train_loss: 1.7612, valid_loss: 1.3224

Epoch [1946/3000], train_loss: 1.7628, valid_loss: 1.4386

Epoch [1947/3000], train_loss: 1.9100, valid_loss: 1.3986

Epoch [1948/3000], train_loss: 2.0625, valid_loss: 1.3576

Epoch [1949/3000], train_loss: 1.7020, valid_loss: 1.3567

Epoch [1950/3000], train_loss: 1.7489, valid_loss: 1.4686

Epoch [1951/3000], train_loss: 1.8333, valid_loss: 1.6915

Epoch [1952/3000], train_loss: 2.1419, valid_loss: 1.2057

Epoch [1953/3000], train_loss: 1.9808, valid_loss: 1.1650

Epoch [1954/3000], train_loss: 2.1664, valid_loss: 1.5945

Epoch [1955/3000], train_loss: 1.9269, valid_loss: 1.3894

Epoch [1956/3000], train_loss: 1.9809, valid_loss: 1.7246

Epoch [1957/3000], train_loss: 1.8775, valid_loss: 1.1406

Epoch [1958/3000], train_loss: 1.8892, valid_loss: 1.0814
Save model at epoch 1958 with loss 1.0814

Epoch [1959/3000], train_loss: 1.7739, valid_loss: 1.1279

Epoch [1960/3000], train_loss: 1.8627, valid_loss: 1.5562

Epoch [1961/3000], train_loss: 1.9377, valid_loss: 1.7084

Epoch [1962/3000], train_loss: 1.9436, valid_loss: 1.4465

Epoch [1963/3000], train_loss: 1.8109, valid_loss: 1.2357

Epoch [1964/3000], train_loss: 2.4035, valid_loss: 1.1394

Epoch [1965/3000], train_loss: 2.0342, valid_loss: 1.5179

Epoch [1966/3000], train_loss: 2.1601, valid_loss: 1.5748

Epoch [1967/3000], train_loss: 2.1259, valid_loss: 1.6709

Epoch [1968/3000], train_loss: 1.8557, valid_loss: 1.3384

Epoch [1969/3000], train_loss: 1.9266, valid_loss: 1.4287

Epoch [1970/3000], train_loss: 1.8805, valid_loss: 1.1981

Epoch [1971/3000], train_loss: 1.9065, valid_loss: 1.2965

Epoch [1972/3000], train_loss: 1.9182, valid_loss: 1.2599

Epoch [1973/3000], train_loss: 1.9614, valid_loss: 1.3463

Epoch [1974/3000], train_loss: 2.0638, valid_loss: 1.2354

Epoch [1975/3000], train_loss: 1.8209, valid_loss: 1.6259

Epoch [1976/3000], train_loss: 1.8610, valid_loss: 1.4595

Epoch [1977/3000], train_loss: 1.8654, valid_loss: 1.4862

Epoch [1978/3000], train_loss: 2.2305, valid_loss: 1.2044

Epoch [1979/3000], train_loss: 2.1815, valid_loss: 1.3219

Epoch [1980/3000], train_loss: 2.0053, valid_loss: 1.3038

Epoch [1981/3000], train_loss: 1.8320, valid_loss: 1.3117

Epoch [1982/3000], train_loss: 1.9976, valid_loss: 1.1458

Epoch [1983/3000], train_loss: 1.9429, valid_loss: 1.4748

Epoch [1984/3000], train_loss: 2.0479, valid_loss: 1.5782

Epoch [1985/3000], train_loss: 1.8429, valid_loss: 1.4981

Epoch [1986/3000], train_loss: 2.1978, valid_loss: 1.2540

Epoch [1987/3000], train_loss: 1.9910, valid_loss: 1.4496

Epoch [1988/3000], train_loss: 2.0265, valid_loss: 1.6666

Epoch [1989/3000], train_loss: 1.7579, valid_loss: 1.6796

Epoch [1990/3000], train_loss: 1.9079, valid_loss: 1.9790

Epoch [1991/3000], train_loss: 2.0522, valid_loss: 1.3590

Epoch [1992/3000], train_loss: 2.0628, valid_loss: 1.5401

Epoch [1993/3000], train_loss: 2.0967, valid_loss: 1.2951

Epoch [1994/3000], train_loss: 1.9445, valid_loss: 1.2232

Epoch [1995/3000], train_loss: 2.2394, valid_loss: 1.3910

Epoch [1996/3000], train_loss: 1.7955, valid_loss: 1.1294

Epoch [1997/3000], train_loss: 2.0024, valid_loss: 1.5163

Epoch [1998/3000], train_loss: 2.1191, valid_loss: 1.1801

Epoch [1999/3000], train_loss: 1.9030, valid_loss: 1.6245

Epoch [2000/3000], train_loss: 2.3653, valid_loss: 1.5378

Epoch [2001/3000], train_loss: 1.9415, valid_loss: 1.4484

Epoch [2002/3000], train_loss: 2.0040, valid_loss: 1.5095

Epoch [2003/3000], train_loss: 1.9401, valid_loss: 1.5656

Epoch [2004/3000], train_loss: 1.7612, valid_loss: 1.3649

Epoch [2005/3000], train_loss: 2.0179, valid_loss: 1.4649

Epoch [2006/3000], train_loss: 1.8463, valid_loss: 1.3551

Epoch [2007/3000], train_loss: 2.1615, valid_loss: 1.7411

Epoch [2008/3000], train_loss: 1.8633, valid_loss: 1.4296

Epoch [2009/3000], train_loss: 2.0067, valid_loss: 1.5458

Epoch [2010/3000], train_loss: 1.8124, valid_loss: 1.3646

Epoch [2011/3000], train_loss: 1.8422, valid_loss: 1.2717

Epoch [2012/3000], train_loss: 1.8558, valid_loss: 1.2837

Epoch [2013/3000], train_loss: 1.8861, valid_loss: 1.1745

Epoch [2014/3000], train_loss: 1.7662, valid_loss: 1.3028

Epoch [2015/3000], train_loss: 1.9640, valid_loss: 1.5603

Epoch [2016/3000], train_loss: 1.9186, valid_loss: 1.3827

Epoch [2017/3000], train_loss: 1.9465, valid_loss: 1.5674

Epoch [2018/3000], train_loss: 1.8637, valid_loss: 1.5278

Epoch [2019/3000], train_loss: 2.0847, valid_loss: 1.2150

Epoch [2020/3000], train_loss: 1.7008, valid_loss: 1.3865

Epoch [2021/3000], train_loss: 1.9331, valid_loss: 1.3737

Epoch [2022/3000], train_loss: 1.8208, valid_loss: 1.6353

Epoch [2023/3000], train_loss: 1.7684, valid_loss: 1.6216

Epoch [2024/3000], train_loss: 2.0072, valid_loss: 1.3865

Epoch [2025/3000], train_loss: 2.0185, valid_loss: 1.1846

Epoch [2026/3000], train_loss: 1.7503, valid_loss: 1.2807

Epoch [2027/3000], train_loss: 1.7753, valid_loss: 1.2814

Epoch [2028/3000], train_loss: 1.8894, valid_loss: 1.4067

Epoch [2029/3000], train_loss: 1.7172, valid_loss: 1.5098

Epoch [2030/3000], train_loss: 2.3176, valid_loss: 1.2710

Epoch [2031/3000], train_loss: 1.7897, valid_loss: 1.0739
Save model at epoch 2031 with loss 1.0739

Epoch [2032/3000], train_loss: 1.8606, valid_loss: 1.2900

Epoch [2033/3000], train_loss: 1.8926, valid_loss: 1.4970

Epoch [2034/3000], train_loss: 1.6790, valid_loss: 1.3717

Epoch [2035/3000], train_loss: 1.9031, valid_loss: 1.2540

Epoch [2036/3000], train_loss: 1.9290, valid_loss: 1.6225

Epoch [2037/3000], train_loss: 1.7692, valid_loss: 1.3455

Epoch [2038/3000], train_loss: 1.8084, valid_loss: 1.1722

Epoch [2039/3000], train_loss: 1.9958, valid_loss: 1.5576

Epoch [2040/3000], train_loss: 2.0147, valid_loss: 1.3596

Epoch [2041/3000], train_loss: 1.9958, valid_loss: 1.6893

Epoch [2042/3000], train_loss: 1.9286, valid_loss: 1.6235

Epoch [2043/3000], train_loss: 2.0315, valid_loss: 1.2751

Epoch [2044/3000], train_loss: 2.1505, valid_loss: 1.2162

Epoch [2045/3000], train_loss: 2.0708, valid_loss: 1.3771

Epoch [2046/3000], train_loss: 1.9192, valid_loss: 1.6364

Epoch [2047/3000], train_loss: 1.7619, valid_loss: 1.3399

Epoch [2048/3000], train_loss: 1.9973, valid_loss: 1.4279

Epoch [2049/3000], train_loss: 1.7260, valid_loss: 1.3924

Epoch [2050/3000], train_loss: 1.8316, valid_loss: 1.4504

Epoch [2051/3000], train_loss: 1.9518, valid_loss: 1.3192

Epoch [2052/3000], train_loss: 1.9166, valid_loss: 1.1942

Epoch [2053/3000], train_loss: 1.8759, valid_loss: 1.3409

Epoch [2054/3000], train_loss: 2.0835, valid_loss: 1.2249

Epoch [2055/3000], train_loss: 1.8383, valid_loss: 1.3227

Epoch [2056/3000], train_loss: 1.8793, valid_loss: 1.2359

Epoch [2057/3000], train_loss: 1.9944, valid_loss: 1.4107

Epoch [2058/3000], train_loss: 1.8193, valid_loss: 1.2830

Epoch [2059/3000], train_loss: 1.8837, valid_loss: 1.4244

Epoch [2060/3000], train_loss: 1.9327, valid_loss: 1.2661

Epoch [2061/3000], train_loss: 2.1602, valid_loss: 1.3259

Epoch [2062/3000], train_loss: 1.8476, valid_loss: 1.4101

Epoch [2063/3000], train_loss: 1.9446, valid_loss: 1.5217

Epoch [2064/3000], train_loss: 1.8516, valid_loss: 1.5813

Epoch [2065/3000], train_loss: 1.9732, valid_loss: 1.2636

Epoch [2066/3000], train_loss: 1.9876, valid_loss: 1.3446

Epoch [2067/3000], train_loss: 1.7633, valid_loss: 1.8481

Epoch [2068/3000], train_loss: 1.9364, valid_loss: 1.4014

Epoch [2069/3000], train_loss: 1.7709, valid_loss: 1.4595

Epoch [2070/3000], train_loss: 1.9186, valid_loss: 1.2427

Epoch [2071/3000], train_loss: 1.8210, valid_loss: 1.4680

Epoch [2072/3000], train_loss: 1.9383, valid_loss: 1.7257

Epoch [2073/3000], train_loss: 1.9638, valid_loss: 1.3048

Epoch [2074/3000], train_loss: 2.0758, valid_loss: 1.2318

Epoch [2075/3000], train_loss: 1.8664, valid_loss: 1.1969

Epoch [2076/3000], train_loss: 1.9684, valid_loss: 1.2686

Epoch [2077/3000], train_loss: 2.0005, valid_loss: 1.1679

Epoch [2078/3000], train_loss: 1.7461, valid_loss: 1.1544

Epoch [2079/3000], train_loss: 1.8825, valid_loss: 1.5191

Epoch [2080/3000], train_loss: 1.8616, valid_loss: 1.3707

Epoch [2081/3000], train_loss: 1.8981, valid_loss: 1.5178

Epoch [2082/3000], train_loss: 2.1906, valid_loss: 1.1256

Epoch [2083/3000], train_loss: 1.8223, valid_loss: 1.5647

Epoch [2084/3000], train_loss: 1.7345, valid_loss: 1.2107

Epoch [2085/3000], train_loss: 1.8068, valid_loss: 1.2344

Epoch [2086/3000], train_loss: 2.0653, valid_loss: 1.3492

Epoch [2087/3000], train_loss: 1.8971, valid_loss: 1.4577

Epoch [2088/3000], train_loss: 1.7743, valid_loss: 1.4472

Epoch [2089/3000], train_loss: 1.6752, valid_loss: 1.4915

Epoch [2090/3000], train_loss: 1.7083, valid_loss: 1.2914

Epoch [2091/3000], train_loss: 1.9870, valid_loss: 1.1298

Epoch [2092/3000], train_loss: 1.8817, valid_loss: 1.2089

Epoch [2093/3000], train_loss: 1.9844, valid_loss: 1.4434

Epoch [2094/3000], train_loss: 1.7748, valid_loss: 1.7880

Epoch [2095/3000], train_loss: 1.8315, valid_loss: 1.4425

Epoch [2096/3000], train_loss: 2.0066, valid_loss: 1.5482

Epoch [2097/3000], train_loss: 1.7265, valid_loss: 1.7446

Epoch [2098/3000], train_loss: 1.8160, valid_loss: 1.2001

Epoch [2099/3000], train_loss: 2.1414, valid_loss: 1.3313

Epoch [2100/3000], train_loss: 1.8090, valid_loss: 1.3692

Epoch [2101/3000], train_loss: 1.8415, valid_loss: 1.4120

Epoch [2102/3000], train_loss: 1.9154, valid_loss: 1.2943

Epoch [2103/3000], train_loss: 1.9349, valid_loss: 1.2905

Epoch [2104/3000], train_loss: 1.8952, valid_loss: 1.6612

Epoch [2105/3000], train_loss: 1.9599, valid_loss: 1.2802

Epoch [2106/3000], train_loss: 1.7853, valid_loss: 1.2017

Epoch [2107/3000], train_loss: 1.8855, valid_loss: 1.3507

Epoch [2108/3000], train_loss: 1.8086, valid_loss: 1.4687

Epoch [2109/3000], train_loss: 1.7252, valid_loss: 1.4082

Epoch [2110/3000], train_loss: 2.0983, valid_loss: 1.3997

Epoch [2111/3000], train_loss: 2.1237, valid_loss: 1.3053

Epoch [2112/3000], train_loss: 1.7476, valid_loss: 1.5008

Epoch [2113/3000], train_loss: 1.9016, valid_loss: 1.5035

Epoch [2114/3000], train_loss: 1.8444, valid_loss: 1.3361

Epoch [2115/3000], train_loss: 2.0281, valid_loss: 1.2630

Epoch [2116/3000], train_loss: 1.8525, valid_loss: 1.1986

Epoch [2117/3000], train_loss: 1.8802, valid_loss: 1.7197

Epoch [2118/3000], train_loss: 1.8788, valid_loss: 1.5541

Epoch [2119/3000], train_loss: 1.6505, valid_loss: 1.6015

Epoch [2120/3000], train_loss: 2.0576, valid_loss: 1.5577

Epoch [2121/3000], train_loss: 1.8717, valid_loss: 1.2509

Epoch [2122/3000], train_loss: 1.8791, valid_loss: 1.2020

Epoch [2123/3000], train_loss: 1.7746, valid_loss: 1.2381

Epoch [2124/3000], train_loss: 1.8026, valid_loss: 1.5934

Epoch [2125/3000], train_loss: 1.9969, valid_loss: 1.4024

Epoch [2126/3000], train_loss: 2.2898, valid_loss: 1.3397

Epoch [2127/3000], train_loss: 2.0159, valid_loss: 1.5555

Epoch [2128/3000], train_loss: 1.6750, valid_loss: 1.4926

Epoch [2129/3000], train_loss: 1.7175, valid_loss: 1.3147

Epoch [2130/3000], train_loss: 1.6674, valid_loss: 1.2688

Epoch [2131/3000], train_loss: 1.7051, valid_loss: 1.2360

Epoch [2132/3000], train_loss: 1.9632, valid_loss: 1.3008

Epoch [2133/3000], train_loss: 1.7202, valid_loss: 1.3256

Epoch [2134/3000], train_loss: 1.7013, valid_loss: 1.4343

Epoch [2135/3000], train_loss: 1.7548, valid_loss: 1.4456

Epoch [2136/3000], train_loss: 1.8061, valid_loss: 1.3137

Epoch [2137/3000], train_loss: 2.0546, valid_loss: 1.2122

Epoch [2138/3000], train_loss: 1.8997, valid_loss: 1.5444

Epoch [2139/3000], train_loss: 1.9613, valid_loss: 1.2158

Epoch [2140/3000], train_loss: 1.7924, valid_loss: 1.1381

Epoch [2141/3000], train_loss: 1.7552, valid_loss: 1.2243

Epoch [2142/3000], train_loss: 1.9425, valid_loss: 1.4160

Epoch [2143/3000], train_loss: 1.7287, valid_loss: 1.3494

Epoch [2144/3000], train_loss: 1.6518, valid_loss: 1.2242

Epoch [2145/3000], train_loss: 1.9673, valid_loss: 1.3797

Epoch [2146/3000], train_loss: 1.8603, valid_loss: 1.2327

Epoch [2147/3000], train_loss: 2.0767, valid_loss: 1.2396

Epoch [2148/3000], train_loss: 1.8207, valid_loss: 1.1042

Epoch [2149/3000], train_loss: 1.8356, valid_loss: 1.3608

Epoch [2150/3000], train_loss: 1.9269, valid_loss: 1.9923

Epoch [2151/3000], train_loss: 2.4215, valid_loss: 1.5192

Epoch [2152/3000], train_loss: 1.8110, valid_loss: 1.2581

Epoch [2153/3000], train_loss: 1.6554, valid_loss: 1.4000

Epoch [2154/3000], train_loss: 1.7224, valid_loss: 1.1688

Epoch [2155/3000], train_loss: 1.9967, valid_loss: 1.3401

Epoch [2156/3000], train_loss: 2.0797, valid_loss: 1.2796

Epoch [2157/3000], train_loss: 1.7803, valid_loss: 1.4500

Epoch [2158/3000], train_loss: 1.8656, valid_loss: 1.2641

Epoch [2159/3000], train_loss: 1.9278, valid_loss: 1.3077

Epoch [2160/3000], train_loss: 2.0798, valid_loss: 1.1195

Epoch [2161/3000], train_loss: 2.0229, valid_loss: 1.1738

Epoch [2162/3000], train_loss: 2.0422, valid_loss: 1.8900

Epoch [2163/3000], train_loss: 1.9735, valid_loss: 1.6541

Epoch [2164/3000], train_loss: 1.8499, valid_loss: 1.2831

Epoch [2165/3000], train_loss: 1.9488, valid_loss: 1.3661

Epoch [2166/3000], train_loss: 1.7466, valid_loss: 1.3129

Epoch [2167/3000], train_loss: 1.9939, valid_loss: 1.7047

Epoch [2168/3000], train_loss: 1.7818, valid_loss: 1.3470

Epoch [2169/3000], train_loss: 2.1759, valid_loss: 1.5979

Epoch [2170/3000], train_loss: 1.7600, valid_loss: 1.4419

Epoch [2171/3000], train_loss: 1.8417, valid_loss: 1.4530

Epoch [2172/3000], train_loss: 1.7126, valid_loss: 1.4007

Epoch [2173/3000], train_loss: 1.8930, valid_loss: 1.4552

Epoch [2174/3000], train_loss: 1.7259, valid_loss: 1.3566

Epoch [2175/3000], train_loss: 1.9269, valid_loss: 1.3175

Epoch [2176/3000], train_loss: 1.8484, valid_loss: 1.5273

Epoch [2177/3000], train_loss: 1.8886, valid_loss: 1.5833

Epoch [2178/3000], train_loss: 1.9642, valid_loss: 1.1749

Epoch [2179/3000], train_loss: 1.7442, valid_loss: 1.2134

Epoch [2180/3000], train_loss: 1.9157, valid_loss: 1.3653

Epoch [2181/3000], train_loss: 1.9217, valid_loss: 1.4006

Epoch [2182/3000], train_loss: 1.7882, valid_loss: 1.3781

Epoch [2183/3000], train_loss: 2.0101, valid_loss: 1.3779

Epoch [2184/3000], train_loss: 1.8987, valid_loss: 1.5363

Epoch [2185/3000], train_loss: 1.9275, valid_loss: 1.4655

Epoch [2186/3000], train_loss: 1.9244, valid_loss: 1.2492

Epoch [2187/3000], train_loss: 1.8005, valid_loss: 1.3956

Epoch [2188/3000], train_loss: 1.7112, valid_loss: 1.3344

Epoch [2189/3000], train_loss: 1.9048, valid_loss: 1.3489

Epoch [2190/3000], train_loss: 1.8018, valid_loss: 1.4307

Epoch [2191/3000], train_loss: 1.8670, valid_loss: 1.4317

Epoch [2192/3000], train_loss: 1.9958, valid_loss: 1.3544

Epoch [2193/3000], train_loss: 1.8810, valid_loss: 1.7085

Epoch [2194/3000], train_loss: 1.8925, valid_loss: 1.2021

Epoch [2195/3000], train_loss: 2.2034, valid_loss: 1.3283

Epoch [2196/3000], train_loss: 1.9074, valid_loss: 1.5526

Epoch [2197/3000], train_loss: 1.7493, valid_loss: 1.3935

Epoch [2198/3000], train_loss: 1.7686, valid_loss: 1.2728

Epoch [2199/3000], train_loss: 1.8935, valid_loss: 1.3747

Epoch [2200/3000], train_loss: 1.7805, valid_loss: 1.0848

Epoch [2201/3000], train_loss: 1.8537, valid_loss: 1.4155

Epoch [2202/3000], train_loss: 2.0450, valid_loss: 1.3500

Epoch [2203/3000], train_loss: 1.6635, valid_loss: 1.4939

Epoch [2204/3000], train_loss: 1.7367, valid_loss: 1.2940

Epoch [2205/3000], train_loss: 1.9377, valid_loss: 1.3868

Epoch [2206/3000], train_loss: 1.7823, valid_loss: 1.1468

Epoch [2207/3000], train_loss: 1.9548, valid_loss: 1.3421

Epoch [2208/3000], train_loss: 1.9540, valid_loss: 1.2126

Epoch [2209/3000], train_loss: 1.8333, valid_loss: 1.0242
Save model at epoch 2209 with loss 1.0242

Epoch [2210/3000], train_loss: 1.7867, valid_loss: 1.2553

Epoch [2211/3000], train_loss: 1.7609, valid_loss: 1.1829

Epoch [2212/3000], train_loss: 1.8219, valid_loss: 1.3219

Epoch [2213/3000], train_loss: 1.7769, valid_loss: 1.6359

Epoch [2214/3000], train_loss: 1.9692, valid_loss: 1.2906

Epoch [2215/3000], train_loss: 1.8139, valid_loss: 1.3775

Epoch [2216/3000], train_loss: 1.9406, valid_loss: 1.3691

Epoch [2217/3000], train_loss: 1.8907, valid_loss: 1.4572

Epoch [2218/3000], train_loss: 1.8557, valid_loss: 1.1803

Epoch [2219/3000], train_loss: 1.8835, valid_loss: 1.4783

Epoch [2220/3000], train_loss: 1.7512, valid_loss: 1.5220

Epoch [2221/3000], train_loss: 2.0143, valid_loss: 1.2898

Epoch [2222/3000], train_loss: 1.8486, valid_loss: 1.1483

Epoch [2223/3000], train_loss: 2.0259, valid_loss: 1.4531

Epoch [2224/3000], train_loss: 1.7794, valid_loss: 1.2335

Epoch [2225/3000], train_loss: 1.9224, valid_loss: 1.3724

Epoch [2226/3000], train_loss: 1.8335, valid_loss: 1.8591

Epoch [2227/3000], train_loss: 2.0229, valid_loss: 1.3310

Epoch [2228/3000], train_loss: 2.0876, valid_loss: 1.1513

Epoch [2229/3000], train_loss: 2.0696, valid_loss: 1.3076

Epoch [2230/3000], train_loss: 1.7481, valid_loss: 1.6028

Epoch [2231/3000], train_loss: 1.7200, valid_loss: 1.4621

Epoch [2232/3000], train_loss: 1.7749, valid_loss: 1.2985

Epoch [2233/3000], train_loss: 1.8440, valid_loss: 1.7927

Epoch [2234/3000], train_loss: 2.2374, valid_loss: 1.3145

Epoch [2235/3000], train_loss: 1.8536, valid_loss: 1.3472

Epoch [2236/3000], train_loss: 2.0341, valid_loss: 1.1638

Epoch [2237/3000], train_loss: 1.8827, valid_loss: 1.2085

Epoch [2238/3000], train_loss: 1.8585, valid_loss: 1.3483

Epoch [2239/3000], train_loss: 2.1129, valid_loss: 1.3296

Epoch [2240/3000], train_loss: 2.0837, valid_loss: 1.6516

Epoch [2241/3000], train_loss: 1.8335, valid_loss: 1.2824

Epoch [2242/3000], train_loss: 1.8780, valid_loss: 1.6899

Epoch [2243/3000], train_loss: 1.7495, valid_loss: 1.2371

Epoch [2244/3000], train_loss: 1.8774, valid_loss: 1.4143

Epoch [2245/3000], train_loss: 1.6860, valid_loss: 1.3969

Epoch [2246/3000], train_loss: 2.0180, valid_loss: 1.3279

Epoch [2247/3000], train_loss: 1.9559, valid_loss: 1.1626

Epoch [2248/3000], train_loss: 1.7790, valid_loss: 1.5877

Epoch [2249/3000], train_loss: 1.6948, valid_loss: 1.5342

Epoch [2250/3000], train_loss: 1.7628, valid_loss: 1.6965

Epoch [2251/3000], train_loss: 1.8099, valid_loss: 1.2704

Epoch [2252/3000], train_loss: 1.8123, valid_loss: 1.4377

Epoch [2253/3000], train_loss: 1.9014, valid_loss: 1.2673

Epoch [2254/3000], train_loss: 1.8969, valid_loss: 1.1630

Epoch [2255/3000], train_loss: 1.6937, valid_loss: 1.3277

Epoch [2256/3000], train_loss: 1.8073, valid_loss: 1.4045

Epoch [2257/3000], train_loss: 1.6960, valid_loss: 1.3533

Epoch [2258/3000], train_loss: 1.8113, valid_loss: 1.3521

Epoch [2259/3000], train_loss: 1.7134, valid_loss: 1.4859

Epoch [2260/3000], train_loss: 1.8139, valid_loss: 1.1711

Epoch [2261/3000], train_loss: 1.9083, valid_loss: 1.3923

Epoch [2262/3000], train_loss: 1.7773, valid_loss: 1.5429

Epoch [2263/3000], train_loss: 2.3490, valid_loss: 1.4203

Epoch [2264/3000], train_loss: 1.7895, valid_loss: 1.2801

Epoch [2265/3000], train_loss: 1.6570, valid_loss: 1.0917

Epoch [2266/3000], train_loss: 1.7375, valid_loss: 1.3735

Epoch [2267/3000], train_loss: 1.6366, valid_loss: 1.5426

Epoch [2268/3000], train_loss: 1.7036, valid_loss: 1.3315

Epoch [2269/3000], train_loss: 1.9947, valid_loss: 1.2062

Epoch [2270/3000], train_loss: 1.8510, valid_loss: 1.2144

Epoch [2271/3000], train_loss: 2.0230, valid_loss: 1.3789

Epoch [2272/3000], train_loss: 1.9072, valid_loss: 1.6482

Epoch [2273/3000], train_loss: 1.9432, valid_loss: 1.1786

Epoch [2274/3000], train_loss: 1.9581, valid_loss: 1.2712

Epoch [2275/3000], train_loss: 1.9269, valid_loss: 1.2870

Epoch [2276/3000], train_loss: 1.8942, valid_loss: 1.3656

Epoch [2277/3000], train_loss: 1.8240, valid_loss: 1.2971

Epoch [2278/3000], train_loss: 1.8508, valid_loss: 1.4323

Epoch [2279/3000], train_loss: 2.0152, valid_loss: 1.4802

Epoch [2280/3000], train_loss: 1.6989, valid_loss: 1.2216

Epoch [2281/3000], train_loss: 1.8452, valid_loss: 1.2745

Epoch [2282/3000], train_loss: 1.9316, valid_loss: 1.0802

Epoch [2283/3000], train_loss: 2.0652, valid_loss: 1.2518

Epoch [2284/3000], train_loss: 1.8966, valid_loss: 1.3463

Epoch [2285/3000], train_loss: 1.8301, valid_loss: 1.4760

Epoch [2286/3000], train_loss: 2.1031, valid_loss: 1.4933

Epoch [2287/3000], train_loss: 1.8356, valid_loss: 1.7064

Epoch [2288/3000], train_loss: 1.7747, valid_loss: 1.2997

Epoch [2289/3000], train_loss: 1.7802, valid_loss: 1.1896

Epoch [2290/3000], train_loss: 1.7580, valid_loss: 1.2890

Epoch [2291/3000], train_loss: 1.9475, valid_loss: 1.1958

Epoch [2292/3000], train_loss: 1.7683, valid_loss: 1.4460

Epoch [2293/3000], train_loss: 1.6955, valid_loss: 1.4958

Epoch [2294/3000], train_loss: 1.8246, valid_loss: 1.5935

Epoch [2295/3000], train_loss: 1.8835, valid_loss: 1.3388

Epoch [2296/3000], train_loss: 1.7508, valid_loss: 1.5081

Epoch [2297/3000], train_loss: 1.7488, valid_loss: 1.3992

Epoch [2298/3000], train_loss: 1.7951, valid_loss: 1.1055

Epoch [2299/3000], train_loss: 2.3505, valid_loss: 1.5226

Epoch [2300/3000], train_loss: 1.8119, valid_loss: 1.5424

Epoch [2301/3000], train_loss: 1.8102, valid_loss: 1.2966

Epoch [2302/3000], train_loss: 2.0534, valid_loss: 1.3100

Epoch [2303/3000], train_loss: 2.0812, valid_loss: 1.3594

Epoch [2304/3000], train_loss: 1.8666, valid_loss: 1.3711

Epoch [2305/3000], train_loss: 1.7926, valid_loss: 1.2829

Epoch [2306/3000], train_loss: 2.0216, valid_loss: 1.2920

Epoch [2307/3000], train_loss: 1.9902, valid_loss: 1.3282

Epoch [2308/3000], train_loss: 1.7538, valid_loss: 1.2280

Epoch [2309/3000], train_loss: 1.6714, valid_loss: 1.3829

Epoch [2310/3000], train_loss: 1.8605, valid_loss: 1.5633

Epoch [2311/3000], train_loss: 1.7894, valid_loss: 1.2959

Epoch [2312/3000], train_loss: 1.6996, valid_loss: 1.2458

Epoch [2313/3000], train_loss: 1.8118, valid_loss: 1.1489

Epoch [2314/3000], train_loss: 1.8605, valid_loss: 1.4750

Epoch [2315/3000], train_loss: 1.8457, valid_loss: 1.4255

Epoch [2316/3000], train_loss: 1.6657, valid_loss: 1.2726

Epoch [2317/3000], train_loss: 1.9124, valid_loss: 1.3153

Epoch [2318/3000], train_loss: 1.8309, valid_loss: 1.2190

Epoch [2319/3000], train_loss: 1.6440, valid_loss: 1.3491

Epoch [2320/3000], train_loss: 1.9211, valid_loss: 1.3716

Epoch [2321/3000], train_loss: 1.7223, valid_loss: 1.4017

Epoch [2322/3000], train_loss: 1.7532, valid_loss: 1.3464

Epoch [2323/3000], train_loss: 1.8399, valid_loss: 1.6110

Epoch [2324/3000], train_loss: 1.7735, valid_loss: 1.2635

Epoch [2325/3000], train_loss: 1.8069, valid_loss: 1.2534

Epoch [2326/3000], train_loss: 1.8919, valid_loss: 1.1513

Epoch [2327/3000], train_loss: 2.0042, valid_loss: 1.3492

Epoch [2328/3000], train_loss: 1.8455, valid_loss: 1.2267

Epoch [2329/3000], train_loss: 1.9391, valid_loss: 1.5577

Epoch [2330/3000], train_loss: 1.8513, valid_loss: 1.4140

Epoch [2331/3000], train_loss: 1.8134, valid_loss: 1.3961

Epoch [2332/3000], train_loss: 1.9771, valid_loss: 1.5202

Epoch [2333/3000], train_loss: 1.7873, valid_loss: 1.2867

Epoch [2334/3000], train_loss: 1.7919, valid_loss: 1.3914

Epoch [2335/3000], train_loss: 1.8622, valid_loss: 1.2012

Epoch [2336/3000], train_loss: 1.8032, valid_loss: 1.5836

Epoch [2337/3000], train_loss: 1.7964, valid_loss: 1.4356

Epoch [2338/3000], train_loss: 1.9194, valid_loss: 1.3519

Epoch [2339/3000], train_loss: 1.8316, valid_loss: 1.2111

Epoch [2340/3000], train_loss: 1.6344, valid_loss: 1.1870

Epoch [2341/3000], train_loss: 1.7634, valid_loss: 1.4240

Epoch [2342/3000], train_loss: 1.8435, valid_loss: 1.2050

Epoch [2343/3000], train_loss: 1.9470, valid_loss: 1.3254

Epoch [2344/3000], train_loss: 1.7098, valid_loss: 1.5922

Epoch [2345/3000], train_loss: 1.7561, valid_loss: 1.2652

Epoch [2346/3000], train_loss: 1.7459, valid_loss: 1.3398

Epoch [2347/3000], train_loss: 1.6904, valid_loss: 1.3139

Epoch [2348/3000], train_loss: 1.8037, valid_loss: 1.4573

Epoch [2349/3000], train_loss: 1.7827, valid_loss: 1.3152

Epoch [2350/3000], train_loss: 1.7269, valid_loss: 1.5219

Epoch [2351/3000], train_loss: 2.1163, valid_loss: 1.2112

Epoch [2352/3000], train_loss: 1.8188, valid_loss: 1.3038

Epoch [2353/3000], train_loss: 1.8324, valid_loss: 1.4742

Epoch [2354/3000], train_loss: 1.6364, valid_loss: 1.2459

Epoch [2355/3000], train_loss: 1.9328, valid_loss: 1.3762

Epoch [2356/3000], train_loss: 1.7114, valid_loss: 1.6000

Epoch [2357/3000], train_loss: 1.8066, valid_loss: 1.1167

Epoch [2358/3000], train_loss: 1.8782, valid_loss: 1.2435

Epoch [2359/3000], train_loss: 1.7433, valid_loss: 1.2924

Epoch [2360/3000], train_loss: 1.8109, valid_loss: 1.2647

Epoch [2361/3000], train_loss: 2.1257, valid_loss: 1.0502

Epoch [2362/3000], train_loss: 1.7898, valid_loss: 1.4150

Epoch [2363/3000], train_loss: 1.7074, valid_loss: 1.5014

Epoch [2364/3000], train_loss: 1.8058, valid_loss: 1.1769

Epoch [2365/3000], train_loss: 1.8203, valid_loss: 1.3616

Epoch [2366/3000], train_loss: 1.9099, valid_loss: 1.4860

Epoch [2367/3000], train_loss: 1.5802, valid_loss: 1.2796

Epoch [2368/3000], train_loss: 1.7934, valid_loss: 1.2237

Epoch [2369/3000], train_loss: 1.9199, valid_loss: 1.2041

Epoch [2370/3000], train_loss: 1.9407, valid_loss: 1.4562

Epoch [2371/3000], train_loss: 2.0075, valid_loss: 1.2219

Epoch [2372/3000], train_loss: 1.8343, valid_loss: 1.2040

Epoch [2373/3000], train_loss: 1.7573, valid_loss: 1.4378

Epoch [2374/3000], train_loss: 1.9064, valid_loss: 1.3499

Epoch [2375/3000], train_loss: 1.7309, valid_loss: 1.4547

Epoch [2376/3000], train_loss: 1.7879, valid_loss: 1.2603

Epoch [2377/3000], train_loss: 1.6486, valid_loss: 1.5290

Epoch [2378/3000], train_loss: 1.7504, valid_loss: 1.6786

Epoch [2379/3000], train_loss: 1.7262, valid_loss: 1.4209

Epoch [2380/3000], train_loss: 1.7318, valid_loss: 1.4724

Epoch [2381/3000], train_loss: 1.9347, valid_loss: 1.3595

Epoch [2382/3000], train_loss: 1.7271, valid_loss: 1.4281

Epoch [2383/3000], train_loss: 1.7380, valid_loss: 1.2146

Epoch [2384/3000], train_loss: 1.6726, valid_loss: 1.2415

Epoch [2385/3000], train_loss: 1.7060, valid_loss: 1.3607

Epoch [2386/3000], train_loss: 1.7487, valid_loss: 1.2109

Epoch [2387/3000], train_loss: 1.9215, valid_loss: 1.3473

Epoch [2388/3000], train_loss: 1.8376, valid_loss: 1.6084

Epoch [2389/3000], train_loss: 2.2659, valid_loss: 1.3551

Epoch [2390/3000], train_loss: 1.7533, valid_loss: 1.2331

Epoch [2391/3000], train_loss: 1.9010, valid_loss: 1.3357

Epoch [2392/3000], train_loss: 1.7048, valid_loss: 1.5709

Epoch [2393/3000], train_loss: 1.7011, valid_loss: 1.7420

Epoch [2394/3000], train_loss: 1.7825, valid_loss: 1.3856

Epoch [2395/3000], train_loss: 1.8479, valid_loss: 1.6228

Epoch [2396/3000], train_loss: 1.9515, valid_loss: 1.3606

Epoch [2397/3000], train_loss: 1.7675, valid_loss: 1.1860

Epoch [2398/3000], train_loss: 1.7498, valid_loss: 1.5964

Epoch [2399/3000], train_loss: 1.7024, valid_loss: 1.2857

Epoch [2400/3000], train_loss: 1.8024, valid_loss: 1.2158

Epoch [2401/3000], train_loss: 1.8225, valid_loss: 1.3035

Epoch [2402/3000], train_loss: 1.9051, valid_loss: 1.5026

Epoch [2403/3000], train_loss: 1.8678, valid_loss: 1.2537

Epoch [2404/3000], train_loss: 1.6787, valid_loss: 1.1868

Epoch [2405/3000], train_loss: 1.7373, valid_loss: 1.2164

Epoch [2406/3000], train_loss: 1.8238, valid_loss: 1.1399

Epoch [2407/3000], train_loss: 1.6951, valid_loss: 1.3847

Epoch [2408/3000], train_loss: 1.6790, valid_loss: 1.3177

Epoch [2409/3000], train_loss: 1.8358, valid_loss: 1.6933

Epoch [2410/3000], train_loss: 1.9079, valid_loss: 1.4874

Epoch [2411/3000], train_loss: 2.0393, valid_loss: 1.4965

Epoch [2412/3000], train_loss: 1.8288, valid_loss: 1.2365

Epoch [2413/3000], train_loss: 1.8828, valid_loss: 1.2360

Epoch [2414/3000], train_loss: 1.8014, valid_loss: 1.3934

Epoch [2415/3000], train_loss: 1.8987, valid_loss: 1.7039

Epoch [2416/3000], train_loss: 1.6282, valid_loss: 1.3163

Epoch [2417/3000], train_loss: 1.7662, valid_loss: 1.2571

Epoch [2418/3000], train_loss: 1.7795, valid_loss: 1.5004

Epoch [2419/3000], train_loss: 1.7396, valid_loss: 1.4612

Epoch [2420/3000], train_loss: 1.7753, valid_loss: 1.4999

Epoch [2421/3000], train_loss: 2.0916, valid_loss: 1.2670

Epoch [2422/3000], train_loss: 1.6408, valid_loss: 1.1635

Epoch [2423/3000], train_loss: 1.7210, valid_loss: 1.2751

Epoch [2424/3000], train_loss: 1.7798, valid_loss: 1.4526

Epoch [2425/3000], train_loss: 1.7742, valid_loss: 1.3222

Epoch [2426/3000], train_loss: 1.8964, valid_loss: 1.2688

Epoch [2427/3000], train_loss: 1.8683, valid_loss: 1.5250

Epoch [2428/3000], train_loss: 1.7694, valid_loss: 1.3909

Epoch [2429/3000], train_loss: 1.6350, valid_loss: 1.5691

Epoch [2430/3000], train_loss: 1.6681, valid_loss: 1.4342

Epoch [2431/3000], train_loss: 1.9631, valid_loss: 1.2685

Epoch [2432/3000], train_loss: 1.7862, valid_loss: 1.3880

Epoch [2433/3000], train_loss: 2.0585, valid_loss: 1.6293

Epoch [2434/3000], train_loss: 1.8245, valid_loss: 1.6672

Epoch [2435/3000], train_loss: 1.6394, valid_loss: 1.3707

Epoch [2436/3000], train_loss: 1.9155, valid_loss: 1.3458

Epoch [2437/3000], train_loss: 1.8922, valid_loss: 1.7484

Epoch [2438/3000], train_loss: 1.8134, valid_loss: 1.4285

Epoch [2439/3000], train_loss: 2.1067, valid_loss: 1.4070

Epoch [2440/3000], train_loss: 1.6954, valid_loss: 1.2210

Epoch [2441/3000], train_loss: 2.0651, valid_loss: 1.4610

Epoch [2442/3000], train_loss: 1.9646, valid_loss: 1.5392

Epoch [2443/3000], train_loss: 1.8108, valid_loss: 1.1190

Epoch [2444/3000], train_loss: 1.6985, valid_loss: 1.3327

Epoch [2445/3000], train_loss: 1.8985, valid_loss: 1.0778

Epoch [2446/3000], train_loss: 1.8138, valid_loss: 1.5780

Epoch [2447/3000], train_loss: 1.7873, valid_loss: 1.5866

Epoch [2448/3000], train_loss: 1.6993, valid_loss: 1.2072

Epoch [2449/3000], train_loss: 1.7081, valid_loss: 1.3085

Epoch [2450/3000], train_loss: 2.0577, valid_loss: 1.1317

Epoch [2451/3000], train_loss: 1.7291, valid_loss: 1.1148

Epoch [2452/3000], train_loss: 1.6966, valid_loss: 1.3157

Epoch [2453/3000], train_loss: 2.0507, valid_loss: 1.2868

Epoch [2454/3000], train_loss: 1.7188, valid_loss: 1.3092

Epoch [2455/3000], train_loss: 1.7434, valid_loss: 1.3689

Epoch [2456/3000], train_loss: 1.9157, valid_loss: 1.3770

Epoch [2457/3000], train_loss: 1.8279, valid_loss: 1.1986

Epoch [2458/3000], train_loss: 1.7222, valid_loss: 1.3449

Epoch [2459/3000], train_loss: 1.6656, valid_loss: 1.2992

Epoch [2460/3000], train_loss: 1.8065, valid_loss: 1.2792

Epoch [2461/3000], train_loss: 1.6892, valid_loss: 1.1872

Epoch [2462/3000], train_loss: 1.7800, valid_loss: 1.6769

Epoch [2463/3000], train_loss: 1.9640, valid_loss: 1.2182

Epoch [2464/3000], train_loss: 1.7152, valid_loss: 1.3545

Epoch [2465/3000], train_loss: 1.7972, valid_loss: 1.2884

Epoch [2466/3000], train_loss: 1.7679, valid_loss: 1.6232

Epoch [2467/3000], train_loss: 1.7399, valid_loss: 1.2273

Epoch [2468/3000], train_loss: 1.9510, valid_loss: 1.1408

Epoch [2469/3000], train_loss: 1.8849, valid_loss: 1.4509

Epoch [2470/3000], train_loss: 1.7422, valid_loss: 1.5289

Epoch [2471/3000], train_loss: 1.8288, valid_loss: 1.1401

Epoch [2472/3000], train_loss: 1.7129, valid_loss: 1.5142

Epoch [2473/3000], train_loss: 1.7418, valid_loss: 1.6690

Epoch [2474/3000], train_loss: 1.7281, valid_loss: 1.4781

Epoch [2475/3000], train_loss: 1.9245, valid_loss: 1.3766

Epoch [2476/3000], train_loss: 1.7917, valid_loss: 1.2161

Epoch [2477/3000], train_loss: 1.7375, valid_loss: 1.5346

Epoch [2478/3000], train_loss: 1.7506, valid_loss: 1.5133

Epoch [2479/3000], train_loss: 1.6234, valid_loss: 1.5520

Epoch [2480/3000], train_loss: 1.8565, valid_loss: 1.3404

Epoch [2481/3000], train_loss: 1.7433, valid_loss: 1.4618

Epoch [2482/3000], train_loss: 1.7711, valid_loss: 1.4457

Epoch [2483/3000], train_loss: 1.7585, valid_loss: 1.5870

Epoch [2484/3000], train_loss: 1.7417, valid_loss: 1.3580

Epoch [2485/3000], train_loss: 1.7693, valid_loss: 1.2789

Epoch [2486/3000], train_loss: 1.9641, valid_loss: 1.2824

Epoch [2487/3000], train_loss: 1.6303, valid_loss: 1.1990

Epoch [2488/3000], train_loss: 2.1653, valid_loss: 1.4875

Epoch [2489/3000], train_loss: 1.9464, valid_loss: 1.3699

Epoch [2490/3000], train_loss: 1.8097, valid_loss: 1.2235

Epoch [2491/3000], train_loss: 1.8906, valid_loss: 1.2714

Epoch [2492/3000], train_loss: 2.0696, valid_loss: 1.2759

Epoch [2493/3000], train_loss: 1.8969, valid_loss: 1.1962

Epoch [2494/3000], train_loss: 1.7669, valid_loss: 1.5681

Epoch [2495/3000], train_loss: 1.7932, valid_loss: 1.3221

Epoch [2496/3000], train_loss: 1.7448, valid_loss: 1.3821

Epoch [2497/3000], train_loss: 1.6358, valid_loss: 1.5641

Epoch [2498/3000], train_loss: 2.1117, valid_loss: 1.3583

Epoch [2499/3000], train_loss: 1.6356, valid_loss: 1.2044

Epoch [2500/3000], train_loss: 1.7457, valid_loss: 1.2849

Epoch [2501/3000], train_loss: 1.6989, valid_loss: 1.6605

Epoch [2502/3000], train_loss: 1.8128, valid_loss: 1.0766

Epoch [2503/3000], train_loss: 1.7491, valid_loss: 1.2845

Epoch [2504/3000], train_loss: 1.7888, valid_loss: 1.3815

Epoch [2505/3000], train_loss: 1.7405, valid_loss: 1.2100

Epoch [2506/3000], train_loss: 1.7256, valid_loss: 1.3259

Epoch [2507/3000], train_loss: 1.6868, valid_loss: 1.3802

Epoch [2508/3000], train_loss: 1.9391, valid_loss: 1.3770

Epoch [2509/3000], train_loss: 2.1193, valid_loss: 1.4523

Epoch [2510/3000], train_loss: 1.8640, valid_loss: 1.2998

Epoch [2511/3000], train_loss: 1.8652, valid_loss: 1.4003

Epoch [2512/3000], train_loss: 1.7522, valid_loss: 1.3196

Epoch [2513/3000], train_loss: 1.7452, valid_loss: 1.5260

Epoch [2514/3000], train_loss: 1.6726, valid_loss: 1.5490

Epoch [2515/3000], train_loss: 1.7512, valid_loss: 1.3455

Epoch [2516/3000], train_loss: 1.6816, valid_loss: 1.1955

Epoch [2517/3000], train_loss: 1.8485, valid_loss: 1.1649

Epoch [2518/3000], train_loss: 1.9184, valid_loss: 1.4998

Epoch [2519/3000], train_loss: 1.8305, valid_loss: 1.2279

Epoch [2520/3000], train_loss: 1.6219, valid_loss: 1.3543

Epoch [2521/3000], train_loss: 1.5113, valid_loss: 1.3865

Epoch [2522/3000], train_loss: 1.6903, valid_loss: 1.2130

Epoch [2523/3000], train_loss: 2.1549, valid_loss: 1.4486

Epoch [2524/3000], train_loss: 1.7819, valid_loss: 1.2324

Epoch [2525/3000], train_loss: 1.6198, valid_loss: 1.4187

Epoch [2526/3000], train_loss: 1.6235, valid_loss: 1.7404

Epoch [2527/3000], train_loss: 1.8327, valid_loss: 1.4267

Epoch [2528/3000], train_loss: 1.7250, valid_loss: 1.3393

Epoch [2529/3000], train_loss: 1.7513, valid_loss: 1.2437

Epoch [2530/3000], train_loss: 1.6861, valid_loss: 1.3271

Epoch [2531/3000], train_loss: 1.7828, valid_loss: 1.2902

Epoch [2532/3000], train_loss: 1.7892, valid_loss: 1.5994

Epoch [2533/3000], train_loss: 1.7493, valid_loss: 1.2328

Epoch [2534/3000], train_loss: 1.6247, valid_loss: 1.2577

Epoch [2535/3000], train_loss: 1.7590, valid_loss: 1.2068

Epoch [2536/3000], train_loss: 1.7568, valid_loss: 1.2379

Epoch [2537/3000], train_loss: 1.7440, valid_loss: 1.3597

Epoch [2538/3000], train_loss: 1.7614, valid_loss: 1.3786

Epoch [2539/3000], train_loss: 1.7662, valid_loss: 1.0688

Epoch [2540/3000], train_loss: 1.9291, valid_loss: 1.3501

Epoch [2541/3000], train_loss: 1.5531, valid_loss: 1.3430

Epoch [2542/3000], train_loss: 1.6807, valid_loss: 1.4185

Epoch [2543/3000], train_loss: 1.5353, valid_loss: 1.2334

Epoch [2544/3000], train_loss: 1.6454, valid_loss: 1.2347

Epoch [2545/3000], train_loss: 1.5786, valid_loss: 1.2792

Epoch [2546/3000], train_loss: 1.7074, valid_loss: 1.2886

Epoch [2547/3000], train_loss: 1.6765, valid_loss: 1.2606

Epoch [2548/3000], train_loss: 1.8090, valid_loss: 1.7580

Epoch [2549/3000], train_loss: 1.6400, valid_loss: 1.2445

Epoch [2550/3000], train_loss: 2.1283, valid_loss: 1.2803

Epoch [2551/3000], train_loss: 1.7650, valid_loss: 1.8032

Epoch [2552/3000], train_loss: 1.7442, valid_loss: 1.3090

Epoch [2553/3000], train_loss: 1.8588, valid_loss: 1.2707

Epoch [2554/3000], train_loss: 1.8463, valid_loss: 1.3875

Epoch [2555/3000], train_loss: 1.7839, valid_loss: 1.3216

Epoch [2556/3000], train_loss: 1.8953, valid_loss: 1.2569

Epoch [2557/3000], train_loss: 1.7008, valid_loss: 1.3354

Epoch [2558/3000], train_loss: 1.8047, valid_loss: 1.8818

Epoch [2559/3000], train_loss: 1.7539, valid_loss: 1.1072

Epoch [2560/3000], train_loss: 1.7786, valid_loss: 1.1670

Epoch [2561/3000], train_loss: 1.5792, valid_loss: 1.5404

Epoch [2562/3000], train_loss: 1.7598, valid_loss: 1.5049

Epoch [2563/3000], train_loss: 2.1322, valid_loss: 1.2235

Epoch [2564/3000], train_loss: 1.7697, valid_loss: 1.0550

Epoch [2565/3000], train_loss: 1.7175, valid_loss: 1.2493

Epoch [2566/3000], train_loss: 1.7809, valid_loss: 1.5587

Epoch [2567/3000], train_loss: 1.7690, valid_loss: 1.2903

Epoch [2568/3000], train_loss: 1.7523, valid_loss: 1.3071

Epoch [2569/3000], train_loss: 1.8368, valid_loss: 1.2176

Epoch [2570/3000], train_loss: 1.6268, valid_loss: 1.2773

Epoch [2571/3000], train_loss: 1.7771, valid_loss: 1.3597

Epoch [2572/3000], train_loss: 1.6335, valid_loss: 1.5301

Epoch [2573/3000], train_loss: 1.8163, valid_loss: 1.2322

Epoch [2574/3000], train_loss: 1.9687, valid_loss: 1.1134

Epoch [2575/3000], train_loss: 1.7030, valid_loss: 1.1867

Epoch [2576/3000], train_loss: 1.7362, valid_loss: 1.3901

Epoch [2577/3000], train_loss: 1.7563, valid_loss: 1.3352

Epoch [2578/3000], train_loss: 1.6757, valid_loss: 1.5400

Epoch [2579/3000], train_loss: 1.7239, valid_loss: 1.1835

Epoch [2580/3000], train_loss: 1.7153, valid_loss: 1.1788

Epoch [2581/3000], train_loss: 1.8399, valid_loss: 1.4014

Epoch [2582/3000], train_loss: 1.7523, valid_loss: 1.3148

Epoch [2583/3000], train_loss: 1.7595, valid_loss: 1.3309

Epoch [2584/3000], train_loss: 1.7648, valid_loss: 1.3125

Epoch [2585/3000], train_loss: 1.7603, valid_loss: 1.6898

Epoch [2586/3000], train_loss: 1.8481, valid_loss: 1.4353

Epoch [2587/3000], train_loss: 1.5126, valid_loss: 1.3096

Epoch [2588/3000], train_loss: 1.8942, valid_loss: 1.1570

Epoch [2589/3000], train_loss: 1.8562, valid_loss: 1.6272

Epoch [2590/3000], train_loss: 1.6067, valid_loss: 1.2389

Epoch [2591/3000], train_loss: 1.5737, valid_loss: 1.2763

Epoch [2592/3000], train_loss: 1.8379, valid_loss: 1.4032

Epoch [2593/3000], train_loss: 1.8061, valid_loss: 1.7126

Epoch [2594/3000], train_loss: 2.0936, valid_loss: 1.4975

Epoch [2595/3000], train_loss: 1.9034, valid_loss: 1.1147

Epoch [2596/3000], train_loss: 1.6842, valid_loss: 1.6645

Epoch [2597/3000], train_loss: 1.9438, valid_loss: 1.4555

Epoch [2598/3000], train_loss: 2.1666, valid_loss: 1.2281

Epoch [2599/3000], train_loss: 1.6534, valid_loss: 1.0906

Epoch [2600/3000], train_loss: 1.7009, valid_loss: 1.3672

Epoch [2601/3000], train_loss: 1.9628, valid_loss: 1.6477

Epoch [2602/3000], train_loss: 1.7040, valid_loss: 1.4050

Epoch [2603/3000], train_loss: 1.8306, valid_loss: 1.5250

Epoch [2604/3000], train_loss: 1.7187, valid_loss: 1.3855

Epoch [2605/3000], train_loss: 1.7851, valid_loss: 1.5252

Epoch [2606/3000], train_loss: 1.4791, valid_loss: 1.1864

Epoch [2607/3000], train_loss: 1.5487, valid_loss: 1.1772

Epoch [2608/3000], train_loss: 1.7730, valid_loss: 1.5462

Epoch [2609/3000], train_loss: 1.6605, valid_loss: 1.4423

Epoch [2610/3000], train_loss: 1.7499, valid_loss: 1.6010
Early stop at epoch 2610

代码

文本

5. 测试

代码

文本

[12]

def save_pred(preds, file):

with open(file, "w", newline="") as f:

writer = csv.writer(f)

writer.writerow(["id", "tested_positive"])

for i, p in enumerate(preds):

writer.writerow([i, p])

model = MyModel(input_dim=x_train.shape[1]).to(device)

model.load_state_dict(torch.load(os.path.join(config["model_dir"], "best_model.pth")))

preds = predict(test_loader, model, device)

save_pred(preds, "pred.csv")

/tmp/ipykernel_97/2294147625.py:9: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
  model.load_state_dict(torch.load(os.path.join(config["model_dir"], "best_model.pth")))

代码

文本

Machine Learning

点个赞吧