准确率与参数设置

[whateverud]

您好，我在运行mr这个数据集，准确率无法达到文中那么高，最高也是74，默认参数甚至更低，请问可以分享一下您的参数设置吗?
环境配置：
Python 3.6.13
Tensorflow 1.12.0
Scipy 1.5.4
参数设置：
learning_rate 0.005
epochs 300
batch_size 128
input_dim 300
Hidden 128
steps 1
dropout 0.5
weight_decay 0
early_stopping -1
max_degree 3
运行日志：
(tensorflow) D:\Python\text-ing1>python build_graph.py mr
using default window size = 3
using default unweighted graph
loading raw data
building graphs for training
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5879/5879 [00:05<00:00, 1087.83it/s]
building graphs for training + validation
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 6532/6532 [00:04<00:00, 1564.62it/s]
building graphs for test
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2568/2568 [00:01<00:00, 1653.97it/s]
max_doc_length 340 min_doc_length 4 average 44.73
training_vocab 8695 test_vocab 7467 intersection 7270

(tensorflow) D:\Python\text-ing1>python train.py --dataset mr
D:\Python\anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\framework\dtypes.py:523: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint8 = np.dtype([("qint8", np.int8, 1)])
D:\Python\anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\framework\dtypes.py:524: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint8 = np.dtype([("quint8", np.uint8, 1)])
D:\Python\anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\framework\dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint16 = np.dtype([("qint16", np.int16, 1)])
D:\Python\anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\framework\dtypes.py:526: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint16 = np.dtype([("quint16", np.uint16, 1)])
D:\Python\anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\framework\dtypes.py:527: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint32 = np.dtype([("qint32", np.int32, 1)])
D:\Python\anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\framework\dtypes.py:532: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
np_resource = np.dtype([("resource", np.ubyte, 1)])
D:\Python\text-ing1\utils.py:82: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray
train_adj = np.array(train_adj)
D:\Python\text-ing1\utils.py:83: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray
val_adj = np.array(val_adj)
D:\Python\text-ing1\utils.py:84: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray
test_adj = np.array(test_adj)
D:\Python\text-ing1\utils.py:85: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray
train_embed = np.array(train_embed)
D:\Python\text-ing1\utils.py:86: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray
val_embed = np.array(val_embed)
D:\Python\text-ing1\utils.py:87: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray
test_embed = np.array(test_embed)
loading training set
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 6398/6398 [00:00<00:00, 9130.35it/s]
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 6398/6398 [00:00<00:00, 9834.68it/s]
loading validation set
100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 710/710 [00:00<00:00, 8878.95it/s]
100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 710/710 [00:00<00:00, 10148.16it/s]
loading test set
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3554/3554 [00:00<00:00, 9547.53it/s]
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3554/3554 [00:00<00:00, 9845.77it/s]
build...
WARNING:tensorflow:From D:\Python\text-ing1\metrics.py:6: softmax_cross_entropy_with_logits (from tensorflow.python.ops.nn_ops) is deprecated and will be removed in a future version.
Instructions for updating:

Future major versions of TensorFlow will allow gradients to flow
into the labels input on backprop by default.

See tf.nn.softmax_cross_entropy_with_logits_v2.

2021-05-09 11:32:07.199176: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
train start...
Epoch: 0000 train_loss= 0.69575 train_acc= 0.51094 val_loss= 0.69226 val_acc= 0.50000 test_acc= 0.50028 time= 10.06635
Epoch: 0001 train_loss= 0.69177 train_acc= 0.52563 val_loss= 0.68898 val_acc= 0.53239 test_acc= 0.55262 time= 9.37085
Epoch: 0002 train_loss= 0.68322 train_acc= 0.56268 val_loss= 0.67658 val_acc= 0.57746 test_acc= 0.58244 time= 9.41362
Epoch: 0003 train_loss= 0.66895 train_acc= 0.58815 val_loss= 0.66829 val_acc= 0.59859 test_acc= 0.60242 time= 9.41338
Epoch: 0004 train_loss= 0.66522 train_acc= 0.59409 val_loss= 0.69396 val_acc= 0.50563 test_acc= 0.50647 time= 9.43484
Epoch: 0005 train_loss= 0.66729 train_acc= 0.58190 val_loss= 0.66119 val_acc= 0.59155 test_acc= 0.59820 time= 9.36946
Epoch: 0006 train_loss= 0.65636 train_acc= 0.61066 val_loss= 0.66293 val_acc= 0.60000 test_acc= 0.58807 time= 9.52164
Epoch: 0007 train_loss= 0.65120 train_acc= 0.61488 val_loss= 0.66714 val_acc= 0.60986 test_acc= 0.59932 time= 9.53432
Epoch: 0008 train_loss= 0.65217 train_acc= 0.61550 val_loss= 0.66052 val_acc= 0.58873 test_acc= 0.59257 time= 9.43989
Epoch: 0009 train_loss= 0.63813 train_acc= 0.63051 val_loss= 0.64807 val_acc= 0.61408 test_acc= 0.60580 time= 9.46848
Epoch: 0010 train_loss= 0.64021 train_acc= 0.63567 val_loss= 0.64397 val_acc= 0.63239 test_acc= 0.61114 time= 9.39541
Epoch: 0011 train_loss= 0.63666 train_acc= 0.62988 val_loss= 0.65440 val_acc= 0.61972 test_acc= 0.60383 time= 9.45876
Epoch: 0012 train_loss= 0.63583 train_acc= 0.64004 val_loss= 0.65723 val_acc= 0.60845 test_acc= 0.59313 time= 9.48956
Epoch: 0013 train_loss= 0.63290 train_acc= 0.63598 val_loss= 0.62616 val_acc= 0.64085 test_acc= 0.63731 time= 9.57913
Epoch: 0014 train_loss= 0.62451 train_acc= 0.64708 val_loss= 0.63261 val_acc= 0.62113 test_acc= 0.62296 time= 9.42966
Epoch: 0015 train_loss= 0.62469 train_acc= 0.65646 val_loss= 0.63072 val_acc= 0.63099 test_acc= 0.62577 time= 9.48841
Epoch: 0016 train_loss= 0.61896 train_acc= 0.65927 val_loss= 0.62247 val_acc= 0.64507 test_acc= 0.62971 time= 9.36453
Epoch: 0017 train_loss= 0.62090 train_acc= 0.66068 val_loss= 0.63092 val_acc= 0.64507 test_acc= 0.62155 time= 9.44694
Epoch: 0018 train_loss= 0.62048 train_acc= 0.65505 val_loss= 0.61991 val_acc= 0.64648 test_acc= 0.63900 time= 9.50112
Epoch: 0019 train_loss= 0.62190 train_acc= 0.65442 val_loss= 0.62285 val_acc= 0.64085 test_acc= 0.64856 time= 9.43798
Epoch: 0020 train_loss= 0.61083 train_acc= 0.66599 val_loss= 0.62599 val_acc= 0.64507 test_acc= 0.63393 time= 9.43447
Epoch: 0021 train_loss= 0.60688 train_acc= 0.67115 val_loss= 0.62512 val_acc= 0.64366 test_acc= 0.63056 time= 9.45628
Epoch: 0022 train_loss= 0.60888 train_acc= 0.66005 val_loss= 0.62133 val_acc= 0.63521 test_acc= 0.64434 time= 9.69254
Epoch: 0023 train_loss= 0.59996 train_acc= 0.67521 val_loss= 0.61779 val_acc= 0.64789 test_acc= 0.63337 time= 9.93541
Epoch: 0024 train_loss= 0.59912 train_acc= 0.66943 val_loss= 0.60914 val_acc= 0.66338 test_acc= 0.64744 time= 9.59789
Epoch: 0025 train_loss= 0.60301 train_acc= 0.67709 val_loss= 0.61318 val_acc= 0.65070 test_acc= 0.63928 time= 9.61154
Epoch: 0026 train_loss= 0.60057 train_acc= 0.67240 val_loss= 0.63171 val_acc= 0.63099 test_acc= 0.62774 time= 9.48291
Epoch: 0027 train_loss= 0.60607 train_acc= 0.66802 val_loss= 0.61070 val_acc= 0.67324 test_acc= 0.65335 time= 9.44387
Epoch: 0028 train_loss= 0.60064 train_acc= 0.67631 val_loss= 0.61669 val_acc= 0.64225 test_acc= 0.63562 time= 9.32728
Epoch: 0029 train_loss= 0.59627 train_acc= 0.68037 val_loss= 0.62801 val_acc= 0.62817 test_acc= 0.62324 time= 9.49590
Epoch: 0030 train_loss= 0.60187 train_acc= 0.66443 val_loss= 0.61688 val_acc= 0.64789 test_acc= 0.63928 time= 9.39059
Epoch: 0031 train_loss= 0.59325 train_acc= 0.68162 val_loss= 0.60755 val_acc= 0.66620 test_acc= 0.65363 time= 9.46285
Epoch: 0032 train_loss= 0.59154 train_acc= 0.67865 val_loss= 0.59926 val_acc= 0.66479 test_acc= 0.66376 time= 9.43294
Epoch: 0033 train_loss= 0.59496 train_acc= 0.67505 val_loss= 0.60086 val_acc= 0.67042 test_acc= 0.66179 time= 9.31140
Epoch: 0034 train_loss= 0.58943 train_acc= 0.67896 val_loss= 0.61667 val_acc= 0.66479 test_acc= 0.63759 time= 9.38982
Epoch: 0035 train_loss= 0.59080 train_acc= 0.68568 val_loss= 0.59542 val_acc= 0.67465 test_acc= 0.65279 time= 9.42639
Epoch: 0036 train_loss= 0.58282 train_acc= 0.69209 val_loss= 0.61697 val_acc= 0.65493 test_acc= 0.63928 time= 9.38227
Epoch: 0037 train_loss= 0.58181 train_acc= 0.68115 val_loss= 0.60567 val_acc= 0.66056 test_acc= 0.64688 time= 9.40806
Epoch: 0038 train_loss= 0.57703 train_acc= 0.69866 val_loss= 0.58810 val_acc= 0.69014 test_acc= 0.66038 time= 9.41947
Epoch: 0039 train_loss= 0.57983 train_acc= 0.69006 val_loss= 0.58503 val_acc= 0.68310 test_acc= 0.66545 time= 9.42167
Epoch: 0040 train_loss= 0.57052 train_acc= 0.70038 val_loss= 0.58115 val_acc= 0.68169 test_acc= 0.66826 time= 9.34553
Epoch: 0041 train_loss= 0.57329 train_acc= 0.69459 val_loss= 0.59113 val_acc= 0.68451 test_acc= 0.66207 time= 9.35682
Epoch: 0042 train_loss= 0.57334 train_acc= 0.69303 val_loss= 0.59184 val_acc= 0.68310 test_acc= 0.65250 time= 9.41156
Epoch: 0043 train_loss= 0.57295 train_acc= 0.70522 val_loss= 0.58733 val_acc= 0.68028 test_acc= 0.65447 time= 9.38891
Epoch: 0044 train_loss= 0.56855 train_acc= 0.69928 val_loss= 0.58302 val_acc= 0.70000 test_acc= 0.66517 time= 9.46548
Epoch: 0045 train_loss= 0.57541 train_acc= 0.69162 val_loss= 0.58648 val_acc= 0.67465 test_acc= 0.66854 time= 9.56229
Epoch: 0046 train_loss= 0.56680 train_acc= 0.70678 val_loss= 0.58039 val_acc= 0.68310 test_acc= 0.67107 time= 9.53400
Epoch: 0047 train_loss= 0.56518 train_acc= 0.69928 val_loss= 0.59025 val_acc= 0.67887 test_acc= 0.65701 time= 9.57426
Epoch: 0048 train_loss= 0.55901 train_acc= 0.71085 val_loss= 0.57852 val_acc= 0.69014 test_acc= 0.66995 time= 9.49446
Epoch: 0049 train_loss= 0.55809 train_acc= 0.70350 val_loss= 0.58617 val_acc= 0.68028 test_acc= 0.66545 time= 9.50039
Epoch: 0050 train_loss= 0.56116 train_acc= 0.70725 val_loss= 0.61384 val_acc= 0.65775 test_acc= 0.63506 time= 9.37594
Epoch: 0051 train_loss= 0.55758 train_acc= 0.70460 val_loss= 0.56670 val_acc= 0.69577 test_acc= 0.67614 time= 9.44596
Epoch: 0052 train_loss= 0.55347 train_acc= 0.70866 val_loss= 0.59473 val_acc= 0.67465 test_acc= 0.65813 time= 9.35023
Epoch: 0053 train_loss= 0.55434 train_acc= 0.70944 val_loss= 0.61905 val_acc= 0.66338 test_acc= 0.64660 time= 9.45921
Epoch: 0054 train_loss= 0.54661 train_acc= 0.71616 val_loss= 0.57742 val_acc= 0.69296 test_acc= 0.67501 time= 9.36394
Epoch: 0055 train_loss= 0.54844 train_acc= 0.72570 val_loss= 0.58679 val_acc= 0.67746 test_acc= 0.65841 time= 9.39498
Epoch: 0056 train_loss= 0.54415 train_acc= 0.72445 val_loss= 0.59058 val_acc= 0.68028 test_acc= 0.67023 time= 9.48117
Epoch: 0057 train_loss= 0.54069 train_acc= 0.72507 val_loss= 0.56899 val_acc= 0.70282 test_acc= 0.67811 time= 9.39584
Epoch: 0058 train_loss= 0.54571 train_acc= 0.72351 val_loss= 0.56660 val_acc= 0.69014 test_acc= 0.67642 time= 9.41704
Epoch: 0059 train_loss= 0.54001 train_acc= 0.72570 val_loss= 0.57387 val_acc= 0.68451 test_acc= 0.67304 time= 9.44571
Epoch: 0060 train_loss= 0.54911 train_acc= 0.71569 val_loss= 0.57932 val_acc= 0.67746 test_acc= 0.66348 time= 9.45830
Epoch: 0061 train_loss= 0.53369 train_acc= 0.73398 val_loss= 0.56237 val_acc= 0.71408 test_acc= 0.68796 time= 9.42036
Epoch: 0062 train_loss= 0.53597 train_acc= 0.73038 val_loss= 0.56399 val_acc= 0.71127 test_acc= 0.68261 time= 9.42772
Epoch: 0063 train_loss= 0.53227 train_acc= 0.73148 val_loss= 0.57755 val_acc= 0.69859 test_acc= 0.66939 time= 9.52330
Epoch: 0064 train_loss= 0.51820 train_acc= 0.73601 val_loss= 0.56583 val_acc= 0.70986 test_acc= 0.68627 time= 9.43127
Epoch: 0065 train_loss= 0.52492 train_acc= 0.72741 val_loss= 0.57010 val_acc= 0.69859 test_acc= 0.67839 time= 9.38220
Epoch: 0066 train_loss= 0.52673 train_acc= 0.73257 val_loss= 0.56701 val_acc= 0.70141 test_acc= 0.68486 time= 9.43901
Epoch: 0067 train_loss= 0.53470 train_acc= 0.73304 val_loss= 0.57407 val_acc= 0.70282 test_acc= 0.68092 time= 9.36912
Epoch: 0068 train_loss= 0.52497 train_acc= 0.73070 val_loss= 0.55444 val_acc= 0.71268 test_acc= 0.68965 time= 9.43445
Epoch: 0069 train_loss= 0.52388 train_acc= 0.74117 val_loss= 0.58293 val_acc= 0.68310 test_acc= 0.66798 time= 9.43005
Epoch: 0070 train_loss= 0.52789 train_acc= 0.72945 val_loss= 0.58188 val_acc= 0.68592 test_acc= 0.67276 time= 9.45276
Epoch: 0071 train_loss= 0.52093 train_acc= 0.73726 val_loss= 0.57112 val_acc= 0.70141 test_acc= 0.68064 time= 9.42288
Epoch: 0072 train_loss= 0.51113 train_acc= 0.73961 val_loss= 0.56504 val_acc= 0.69296 test_acc= 0.68542 time= 9.36708
Epoch: 0073 train_loss= 0.51124 train_acc= 0.74992 val_loss= 0.56535 val_acc= 0.69577 test_acc= 0.68768 time= 9.37870
Epoch: 0074 train_loss= 0.51491 train_acc= 0.74258 val_loss= 0.55237 val_acc= 0.71831 test_acc= 0.69555 time= 9.41619
Epoch: 0075 train_loss= 0.50697 train_acc= 0.75133 val_loss= 0.55362 val_acc= 0.71549 test_acc= 0.69527 time= 9.38604
Epoch: 0076 train_loss= 0.51111 train_acc= 0.74773 val_loss= 0.54947 val_acc= 0.71408 test_acc= 0.69837 time= 9.49435
Epoch: 0077 train_loss= 0.50315 train_acc= 0.74992 val_loss= 0.55632 val_acc= 0.71408 test_acc= 0.69977 time= 9.40648
Epoch: 0078 train_loss= 0.50476 train_acc= 0.74648 val_loss= 0.55781 val_acc= 0.71268 test_acc= 0.69162 time= 9.39114
Epoch: 0079 train_loss= 0.50042 train_acc= 0.74977 val_loss= 0.55939 val_acc= 0.71549 test_acc= 0.69696 time= 9.49069
Epoch: 0080 train_loss= 0.49152 train_acc= 0.75711 val_loss= 0.56260 val_acc= 0.71690 test_acc= 0.69612 time= 9.43299
Epoch: 0081 train_loss= 0.50067 train_acc= 0.75430 val_loss= 0.56319 val_acc= 0.70563 test_acc= 0.68993 time= 9.49183
Epoch: 0082 train_loss= 0.48840 train_acc= 0.76008 val_loss= 0.57394 val_acc= 0.70704 test_acc= 0.69415 time= 9.44209
Epoch: 0083 train_loss= 0.48515 train_acc= 0.76164 val_loss= 0.56156 val_acc= 0.71268 test_acc= 0.69471 time= 9.43980
Epoch: 0084 train_loss= 0.48637 train_acc= 0.76196 val_loss= 0.58453 val_acc= 0.70845 test_acc= 0.69443 time= 9.41256
Epoch: 0085 train_loss= 0.47967 train_acc= 0.76493 val_loss= 0.55377 val_acc= 0.72254 test_acc= 0.69949 time= 9.39790
Epoch: 0086 train_loss= 0.48991 train_acc= 0.75367 val_loss= 0.58508 val_acc= 0.70563 test_acc= 0.69246 time= 9.45006
Epoch: 0087 train_loss= 0.48204 train_acc= 0.76211 val_loss= 0.55074 val_acc= 0.71972 test_acc= 0.70090 time= 9.40147
Epoch: 0088 train_loss= 0.47860 train_acc= 0.76383 val_loss= 0.56268 val_acc= 0.71690 test_acc= 0.70343 time= 9.43493
Epoch: 0089 train_loss= 0.47832 train_acc= 0.76571 val_loss= 0.57382 val_acc= 0.70563 test_acc= 0.69049 time= 9.47057
Epoch: 0090 train_loss= 0.47783 train_acc= 0.76586 val_loss= 0.57692 val_acc= 0.69577 test_acc= 0.68824 time= 9.38049
Epoch: 0091 train_loss= 0.47617 train_acc= 0.76633 val_loss= 0.56344 val_acc= 0.71972 test_acc= 0.69921 time= 9.47777
Epoch: 0092 train_loss= 0.47945 train_acc= 0.76790 val_loss= 0.57406 val_acc= 0.70986 test_acc= 0.69668 time= 9.40344
Epoch: 0093 train_loss= 0.47006 train_acc= 0.77102 val_loss= 0.55573 val_acc= 0.71831 test_acc= 0.69977 time= 9.46768
Epoch: 0094 train_loss= 0.46662 train_acc= 0.77259 val_loss= 0.58422 val_acc= 0.69718 test_acc= 0.69865 time= 9.42955
Epoch: 0095 train_loss= 0.46412 train_acc= 0.77384 val_loss= 0.55243 val_acc= 0.70423 test_acc= 0.70934 time= 9.44401
Epoch: 0096 train_loss= 0.46291 train_acc= 0.77602 val_loss= 0.58821 val_acc= 0.71268 test_acc= 0.69302 time= 9.38259
Epoch: 0097 train_loss= 0.45973 train_acc= 0.77165 val_loss= 0.57701 val_acc= 0.70845 test_acc= 0.69133 time= 9.36779
Epoch: 0098 train_loss= 0.45685 train_acc= 0.77931 val_loss= 0.59032 val_acc= 0.70423 test_acc= 0.69724 time= 9.45963
Epoch: 0099 train_loss= 0.45767 train_acc= 0.78056 val_loss= 0.58403 val_acc= 0.70423 test_acc= 0.70597 time= 9.41978
Epoch: 0100 train_loss= 0.46881 train_acc= 0.77243 val_loss= 0.58420 val_acc= 0.70282 test_acc= 0.69246 time= 9.40690
Epoch: 0101 train_loss= 0.45332 train_acc= 0.78650 val_loss= 0.59482 val_acc= 0.71408 test_acc= 0.69105 time= 9.45697
Epoch: 0102 train_loss= 0.46016 train_acc= 0.78118 val_loss= 0.57547 val_acc= 0.71831 test_acc= 0.70850 time= 9.35378
Epoch: 0103 train_loss= 0.45195 train_acc= 0.78321 val_loss= 0.55837 val_acc= 0.70986 test_acc= 0.70653 time= 9.36887
Epoch: 0104 train_loss= 0.44663 train_acc= 0.78446 val_loss= 0.57434 val_acc= 0.70423 test_acc= 0.70174 time= 9.35993
Epoch: 0105 train_loss= 0.44569 train_acc= 0.78618 val_loss= 0.58374 val_acc= 0.71549 test_acc= 0.70146 time= 9.33158
Epoch: 0106 train_loss= 0.44866 train_acc= 0.78587 val_loss= 0.56979 val_acc= 0.72113 test_acc= 0.70371 time= 9.50261
Epoch: 0107 train_loss= 0.42919 train_acc= 0.79337 val_loss= 0.58493 val_acc= 0.72958 test_acc= 0.69893 time= 9.42851
Epoch: 0108 train_loss= 0.44082 train_acc= 0.79431 val_loss= 0.55284 val_acc= 0.72394 test_acc= 0.70203 time= 9.42461
Epoch: 0109 train_loss= 0.42901 train_acc= 0.79884 val_loss= 0.58290 val_acc= 0.72676 test_acc= 0.70568 time= 9.49824
Epoch: 0110 train_loss= 0.43164 train_acc= 0.79415 val_loss= 0.57066 val_acc= 0.72394 test_acc= 0.71019 time= 9.37873
Epoch: 0111 train_loss= 0.43608 train_acc= 0.79619 val_loss= 0.55648 val_acc= 0.73662 test_acc= 0.71384 time= 9.40850
Epoch: 0112 train_loss= 0.43360 train_acc= 0.79462 val_loss= 0.54401 val_acc= 0.73239 test_acc= 0.70709 time= 9.35584
Epoch: 0113 train_loss= 0.43198 train_acc= 0.79697 val_loss= 0.58489 val_acc= 0.71127 test_acc= 0.70625 time= 9.41200
Epoch: 0114 train_loss= 0.42983 train_acc= 0.79384 val_loss= 0.57864 val_acc= 0.72676 test_acc= 0.70287 time= 9.43206
Epoch: 0115 train_loss= 0.40749 train_acc= 0.81197 val_loss= 0.60357 val_acc= 0.71831 test_acc= 0.70597 time= 9.34099
Epoch: 0116 train_loss= 0.41109 train_acc= 0.80416 val_loss= 0.57735 val_acc= 0.72254 test_acc= 0.71075 time= 9.41332
Epoch: 0117 train_loss= 0.41575 train_acc= 0.81041 val_loss= 0.57531 val_acc= 0.73521 test_acc= 0.70878 time= 9.44986
Epoch: 0118 train_loss= 0.42112 train_acc= 0.80119 val_loss= 0.58868 val_acc= 0.72113 test_acc= 0.71019 time= 9.42124
Epoch: 0119 train_loss= 0.41584 train_acc= 0.80994 val_loss= 0.56866 val_acc= 0.72113 test_acc= 0.71300 time= 9.57050
Epoch: 0120 train_loss= 0.41819 train_acc= 0.80494 val_loss= 0.56076 val_acc= 0.73521 test_acc= 0.71412 time= 9.67463
Epoch: 0121 train_loss= 0.41302 train_acc= 0.80416 val_loss= 0.58956 val_acc= 0.71268 test_acc= 0.71047 time= 9.37353
Epoch: 0122 train_loss= 0.42278 train_acc= 0.80275 val_loss= 0.58229 val_acc= 0.71690 test_acc= 0.71328 time= 9.31625
Epoch: 0123 train_loss= 0.41020 train_acc= 0.80807 val_loss= 0.56670 val_acc= 0.71972 test_acc= 0.72032 time= 9.43863
Epoch: 0124 train_loss= 0.39935 train_acc= 0.81166 val_loss= 0.60529 val_acc= 0.70563 test_acc= 0.71609 time= 9.37355
Epoch: 0125 train_loss= 0.40125 train_acc= 0.81260 val_loss= 0.58211 val_acc= 0.71831 test_acc= 0.71497 time= 9.42178
Epoch: 0126 train_loss= 0.39861 train_acc= 0.81166 val_loss= 0.58454 val_acc= 0.72817 test_acc= 0.71384 time= 9.49300
Epoch: 0127 train_loss= 0.39411 train_acc= 0.81666 val_loss= 0.59354 val_acc= 0.72394 test_acc= 0.71384 time= 9.80655
Epoch: 0128 train_loss= 0.40381 train_acc= 0.80744 val_loss= 0.59680 val_acc= 0.71972 test_acc= 0.71750 time= 9.37681
Epoch: 0129 train_loss= 0.39638 train_acc= 0.81713 val_loss= 0.59846 val_acc= 0.71972 test_acc= 0.71806 time= 9.41873
Epoch: 0130 train_loss= 0.39646 train_acc= 0.81682 val_loss= 0.58924 val_acc= 0.72535 test_acc= 0.72032 time= 9.46292
Epoch: 0131 train_loss= 0.39027 train_acc= 0.81947 val_loss= 0.58117 val_acc= 0.73099 test_acc= 0.72003 time= 9.33863
Epoch: 0132 train_loss= 0.38905 train_acc= 0.82448 val_loss= 0.59530 val_acc= 0.73803 test_acc= 0.72032 time= 9.40483
Epoch: 0133 train_loss= 0.39510 train_acc= 0.81229 val_loss= 0.59878 val_acc= 0.72394 test_acc= 0.72032 time= 9.49056
Epoch: 0134 train_loss= 0.38096 train_acc= 0.82495 val_loss= 0.58953 val_acc= 0.72254 test_acc= 0.72257 time= 9.41398
Epoch: 0135 train_loss= 0.38859 train_acc= 0.81947 val_loss= 0.58494 val_acc= 0.71408 test_acc= 0.71525 time= 9.41143
Epoch: 0136 train_loss= 0.39587 train_acc= 0.81760 val_loss= 0.58340 val_acc= 0.72817 test_acc= 0.71638 time= 9.31381
Epoch: 0137 train_loss= 0.37793 train_acc= 0.82510 val_loss= 0.59162 val_acc= 0.72676 test_acc= 0.71891 time= 9.30900
Epoch: 0138 train_loss= 0.37971 train_acc= 0.82776 val_loss= 0.59937 val_acc= 0.73099 test_acc= 0.71750 time= 9.38446
Epoch: 0139 train_loss= 0.37744 train_acc= 0.82917 val_loss= 0.60426 val_acc= 0.72535 test_acc= 0.71750 time= 9.35224
Epoch: 0140 train_loss= 0.37850 train_acc= 0.82791 val_loss= 0.58454 val_acc= 0.72113 test_acc= 0.71947 time= 9.42728
Epoch: 0141 train_loss= 0.37923 train_acc= 0.82323 val_loss= 0.60706 val_acc= 0.71972 test_acc= 0.71778 time= 9.56808
Epoch: 0142 train_loss= 0.37662 train_acc= 0.82526 val_loss= 0.59700 val_acc= 0.72958 test_acc= 0.72032 time= 9.32104
Epoch: 0143 train_loss= 0.36976 train_acc= 0.83323 val_loss= 0.60116 val_acc= 0.72254 test_acc= 0.72060 time= 9.38784
Epoch: 0144 train_loss= 0.36209 train_acc= 0.83589 val_loss= 0.61184 val_acc= 0.72958 test_acc= 0.72116 time= 9.37734
Epoch: 0145 train_loss= 0.37256 train_acc= 0.82979 val_loss= 0.58463 val_acc= 0.72254 test_acc= 0.72397 time= 9.31005
Epoch: 0146 train_loss= 0.36009 train_acc= 0.83886 val_loss= 0.61629 val_acc= 0.72817 test_acc= 0.71835 time= 9.50525
Epoch: 0147 train_loss= 0.37229 train_acc= 0.83088 val_loss= 0.61193 val_acc= 0.72113 test_acc= 0.72088 time= 9.38858
Epoch: 0148 train_loss= 0.36914 train_acc= 0.83104 val_loss= 0.60582 val_acc= 0.73239 test_acc= 0.72257 time= 9.36295
Epoch: 0149 train_loss= 0.36389 train_acc= 0.83651 val_loss= 0.59865 val_acc= 0.71831 test_acc= 0.72482 time= 9.35121
Epoch: 0150 train_loss= 0.36667 train_acc= 0.83385 val_loss= 0.57042 val_acc= 0.73099 test_acc= 0.72341 time= 9.33676
Epoch: 0151 train_loss= 0.35544 train_acc= 0.83807 val_loss= 0.64527 val_acc= 0.70282 test_acc= 0.71553 time= 9.42994
Epoch: 0152 train_loss= 0.36471 train_acc= 0.83276 val_loss= 0.59884 val_acc= 0.72113 test_acc= 0.71919 time= 9.42044
Epoch: 0153 train_loss= 0.35266 train_acc= 0.84417 val_loss= 0.59181 val_acc= 0.73521 test_acc= 0.72341 time= 9.28569
Epoch: 0154 train_loss= 0.36347 train_acc= 0.83323 val_loss= 0.61167 val_acc= 0.72254 test_acc= 0.72735 time= 9.38037
Epoch: 0155 train_loss= 0.36133 train_acc= 0.83854 val_loss= 0.59357 val_acc= 0.72535 test_acc= 0.72622 time= 9.41030
Epoch: 0156 train_loss= 0.34430 train_acc= 0.85308 val_loss= 0.63279 val_acc= 0.71549 test_acc= 0.72228 time= 9.31758
Epoch: 0157 train_loss= 0.35987 train_acc= 0.83698 val_loss= 0.60655 val_acc= 0.72394 test_acc= 0.72988 time= 9.41044
Epoch: 0158 train_loss= 0.35081 train_acc= 0.84386 val_loss= 0.65924 val_acc= 0.70423 test_acc= 0.72088 time= 9.96002
Epoch: 0159 train_loss= 0.35186 train_acc= 0.83854 val_loss= 0.59736 val_acc= 0.71549 test_acc= 0.72847 time= 9.88755
Epoch: 0160 train_loss= 0.34767 train_acc= 0.83886 val_loss= 0.62272 val_acc= 0.72817 test_acc= 0.72622 time= 9.39469
Epoch: 0161 train_loss= 0.34599 train_acc= 0.84620 val_loss= 0.63145 val_acc= 0.72535 test_acc= 0.72369 time= 9.34634
Epoch: 0162 train_loss= 0.34574 train_acc= 0.84745 val_loss= 0.62908 val_acc= 0.72817 test_acc= 0.72819 time= 9.31153
Epoch: 0163 train_loss= 0.33907 train_acc= 0.84745 val_loss= 0.62480 val_acc= 0.73239 test_acc= 0.72819 time= 9.39163
Epoch: 0164 train_loss= 0.34932 train_acc= 0.84323 val_loss= 0.61683 val_acc= 0.72817 test_acc= 0.72707 time= 9.35730
Epoch: 0165 train_loss= 0.33863 train_acc= 0.84745 val_loss= 0.60814 val_acc= 0.72394 test_acc= 0.72960 time= 9.43012
Epoch: 0166 train_loss= 0.33578 train_acc= 0.85183 val_loss= 0.63145 val_acc= 0.72535 test_acc= 0.73298 time= 9.33584
Epoch: 0167 train_loss= 0.33308 train_acc= 0.85058 val_loss= 0.61737 val_acc= 0.73521 test_acc= 0.73129 time= 9.34226
Epoch: 0168 train_loss= 0.34508 train_acc= 0.84448 val_loss= 0.59861 val_acc= 0.72817 test_acc= 0.73382 time= 9.36551
Epoch: 0169 train_loss= 0.33622 train_acc= 0.85027 val_loss= 0.61087 val_acc= 0.72254 test_acc= 0.73101 time= 9.42611
Epoch: 0170 train_loss= 0.32623 train_acc= 0.85277 val_loss= 0.65388 val_acc= 0.71690 test_acc= 0.72566 time= 9.33569
Epoch: 0171 train_loss= 0.32377 train_acc= 0.85605 val_loss= 0.65708 val_acc= 0.73662 test_acc= 0.72904 time= 9.44486
Epoch: 0172 train_loss= 0.33807 train_acc= 0.84448 val_loss= 0.61732 val_acc= 0.73380 test_acc= 0.72707 time= 9.35273
Epoch: 0173 train_loss= 0.33269 train_acc= 0.84948 val_loss= 0.60996 val_acc= 0.73099 test_acc= 0.73016 time= 9.41739
Epoch: 0174 train_loss= 0.32402 train_acc= 0.85417 val_loss= 0.62004 val_acc= 0.73662 test_acc= 0.72904 time= 9.43093
Epoch: 0175 train_loss= 0.32745 train_acc= 0.85214 val_loss= 0.64181 val_acc= 0.71549 test_acc= 0.73523 time= 9.36108
Epoch: 0176 train_loss= 0.32169 train_acc= 0.85886 val_loss= 0.65001 val_acc= 0.72817 test_acc= 0.73157 time= 9.42805
Epoch: 0177 train_loss= 0.31034 train_acc= 0.86480 val_loss= 0.63860 val_acc= 0.74085 test_acc= 0.73016 time= 9.35984
Epoch: 0178 train_loss= 0.33070 train_acc= 0.85058 val_loss= 0.60783 val_acc= 0.72535 test_acc= 0.73298 time= 9.35859
Epoch: 0179 train_loss= 0.31777 train_acc= 0.85824 val_loss= 0.63734 val_acc= 0.73380 test_acc= 0.73101 time= 9.43445
Epoch: 0180 train_loss= 0.31881 train_acc= 0.85777 val_loss= 0.65004 val_acc= 0.73521 test_acc= 0.73326 time= 9.35490
Epoch: 0181 train_loss= 0.32317 train_acc= 0.85917 val_loss= 0.63719 val_acc= 0.74225 test_acc= 0.73185 time= 9.34799
Epoch: 0182 train_loss= 0.32433 train_acc= 0.85714 val_loss= 0.64764 val_acc= 0.73662 test_acc= 0.73270 time= 9.39804
Epoch: 0183 train_loss= 0.31378 train_acc= 0.86105 val_loss= 0.65028 val_acc= 0.73239 test_acc= 0.73073 time= 9.33598
Epoch: 0184 train_loss= 0.31487 train_acc= 0.85964 val_loss= 0.63196 val_acc= 0.73380 test_acc= 0.72735 time= 9.44273
Epoch: 0185 train_loss= 0.30901 train_acc= 0.86183 val_loss= 0.66464 val_acc= 0.73662 test_acc= 0.73438 time= 9.31948
Epoch: 0186 train_loss= 0.31742 train_acc= 0.86199 val_loss= 0.63261 val_acc= 0.72958 test_acc= 0.73607 time= 9.38919
Epoch: 0187 train_loss= 0.30934 train_acc= 0.86590 val_loss= 0.62695 val_acc= 0.73380 test_acc= 0.73213 time= 9.40234
Epoch: 0188 train_loss= 0.31406 train_acc= 0.86277 val_loss= 0.64771 val_acc= 0.74085 test_acc= 0.73185 time= 9.39779
Epoch: 0189 train_loss= 0.30987 train_acc= 0.86261 val_loss= 0.66118 val_acc= 0.72817 test_acc= 0.73016 time= 9.29694
Epoch: 0190 train_loss= 0.30590 train_acc= 0.86793 val_loss= 0.66389 val_acc= 0.73380 test_acc= 0.73129 time= 9.36289
Epoch: 0191 train_loss= 0.31167 train_acc= 0.86433 val_loss= 0.63139 val_acc= 0.73662 test_acc= 0.73438 time= 9.39836
Epoch: 0192 train_loss= 0.29010 train_acc= 0.87293 val_loss= 0.69216 val_acc= 0.73521 test_acc= 0.73382 time= 9.51976
Epoch: 0193 train_loss= 0.29448 train_acc= 0.87480 val_loss= 0.66219 val_acc= 0.73239 test_acc= 0.73354 time= 9.59536
Epoch: 0194 train_loss= 0.30787 train_acc= 0.86465 val_loss= 0.65270 val_acc= 0.73099 test_acc= 0.73213 time= 9.36797
Epoch: 0195 train_loss= 0.30018 train_acc= 0.86730 val_loss= 0.67225 val_acc= 0.73099 test_acc= 0.73354 time= 9.44877
Epoch: 0196 train_loss= 0.30359 train_acc= 0.86918 val_loss= 0.61512 val_acc= 0.73521 test_acc= 0.73917 time= 9.34018
Epoch: 0197 train_loss= 0.30450 train_acc= 0.86824 val_loss= 0.65564 val_acc= 0.72817 test_acc= 0.72988 time= 9.40780
Epoch: 0198 train_loss= 0.29891 train_acc= 0.86715 val_loss= 0.64581 val_acc= 0.72817 test_acc= 0.73382 time= 9.41069
Epoch: 0199 train_loss= 0.29888 train_acc= 0.86418 val_loss= 0.68361 val_acc= 0.73521 test_acc= 0.73889 time= 9.42118
Epoch: 0200 train_loss= 0.30234 train_acc= 0.86808 val_loss= 0.64843 val_acc= 0.73521 test_acc= 0.73438 time= 9.37062
Epoch: 0201 train_loss= 0.30521 train_acc= 0.86652 val_loss= 0.63632 val_acc= 0.74648 test_acc= 0.73270 time= 9.37372
Epoch: 0202 train_loss= 0.29607 train_acc= 0.87027 val_loss= 0.67864 val_acc= 0.73662 test_acc= 0.73298 time= 9.30866
Epoch: 0203 train_loss= 0.30031 train_acc= 0.86668 val_loss= 0.66764 val_acc= 0.72676 test_acc= 0.73270 time= 9.42382
Epoch: 0204 train_loss= 0.29469 train_acc= 0.87168 val_loss= 0.66391 val_acc= 0.71972 test_acc= 0.73326 time= 9.34886
Epoch: 0205 train_loss= 0.29236 train_acc= 0.87449 val_loss= 0.65517 val_acc= 0.73099 test_acc= 0.73016 time= 9.39549
Epoch: 0206 train_loss= 0.28222 train_acc= 0.87934 val_loss= 0.71691 val_acc= 0.72817 test_acc= 0.73101 time= 9.36884
Epoch: 0207 train_loss= 0.28236 train_acc= 0.87699 val_loss= 0.68498 val_acc= 0.72535 test_acc= 0.73129 time= 9.35433
Epoch: 0208 train_loss= 0.29061 train_acc= 0.87777 val_loss= 0.71091 val_acc= 0.72254 test_acc= 0.72904 time= 9.36703
Epoch: 0209 train_loss= 0.29422 train_acc= 0.87355 val_loss= 0.65503 val_acc= 0.73099 test_acc= 0.72988 time= 9.35378
Epoch: 0210 train_loss= 0.28849 train_acc= 0.87480 val_loss= 0.65143 val_acc= 0.73239 test_acc= 0.73410 time= 9.45285
Epoch: 0211 train_loss= 0.28198 train_acc= 0.88199 val_loss= 0.65315 val_acc= 0.73099 test_acc= 0.73579 time= 9.37534
Epoch: 0212 train_loss= 0.29390 train_acc= 0.87246 val_loss= 0.63479 val_acc= 0.74225 test_acc= 0.73241 time= 9.33378
Epoch: 0213 train_loss= 0.28111 train_acc= 0.88262 val_loss= 0.66012 val_acc= 0.73662 test_acc= 0.73044 time= 9.33752
Epoch: 0214 train_loss= 0.28540 train_acc= 0.87527 val_loss= 0.63557 val_acc= 0.73944 test_acc= 0.73663 time= 9.35667
Epoch: 0215 train_loss= 0.29046 train_acc= 0.87168 val_loss= 0.67847 val_acc= 0.73803 test_acc= 0.74029 time= 9.37935
Epoch: 0216 train_loss= 0.27732 train_acc= 0.88278 val_loss= 0.64426 val_acc= 0.74366 test_acc= 0.73354 time= 9.69392
Epoch: 0217 train_loss= 0.28077 train_acc= 0.87668 val_loss= 0.66215 val_acc= 0.73521 test_acc= 0.73467 time= 9.36394
Epoch: 0218 train_loss= 0.27585 train_acc= 0.88074 val_loss= 0.70318 val_acc= 0.73099 test_acc= 0.73467 time= 9.38656
Epoch: 0219 train_loss= 0.27480 train_acc= 0.88184 val_loss= 0.72657 val_acc= 0.72676 test_acc= 0.72791 time= 9.51133
Epoch: 0220 train_loss= 0.27627 train_acc= 0.88184 val_loss= 0.66763 val_acc= 0.72394 test_acc= 0.73748 time= 9.41609
Epoch: 0221 train_loss= 0.28462 train_acc= 0.87527 val_loss= 0.62838 val_acc= 0.74085 test_acc= 0.73354 time= 9.61491
Epoch: 0222 train_loss= 0.28093 train_acc= 0.87777 val_loss= 0.67483 val_acc= 0.73239 test_acc= 0.73044 time= 9.61457
Epoch: 0223 train_loss= 0.28075 train_acc= 0.88028 val_loss= 0.61219 val_acc= 0.74225 test_acc= 0.73073 time= 9.47224
Epoch: 0224 train_loss= 0.26586 train_acc= 0.88981 val_loss= 0.69529 val_acc= 0.73099 test_acc= 0.73044 time= 9.38987
Epoch: 0225 train_loss= 0.28823 train_acc= 0.87512 val_loss= 0.66700 val_acc= 0.72535 test_acc= 0.73073 time= 9.45238
Epoch: 0226 train_loss= 0.26763 train_acc= 0.88512 val_loss= 0.69256 val_acc= 0.73521 test_acc= 0.73270 time= 9.34147
Epoch: 0227 train_loss= 0.26706 train_acc= 0.88731 val_loss= 0.69621 val_acc= 0.72535 test_acc= 0.74198 time= 9.38819
Epoch: 0228 train_loss= 0.27970 train_acc= 0.87981 val_loss= 0.67239 val_acc= 0.71408 test_acc= 0.73776 time= 9.40750
Epoch: 0229 train_loss= 0.28627 train_acc= 0.88121 val_loss= 0.67689 val_acc= 0.72958 test_acc= 0.73241 time= 9.43458
Epoch: 0230 train_loss= 0.27617 train_acc= 0.88512 val_loss= 0.62532 val_acc= 0.74507 test_acc= 0.73213 time= 9.49115
Epoch: 0231 train_loss= 0.26751 train_acc= 0.88621 val_loss= 0.67229 val_acc= 0.72958 test_acc= 0.73635 time= 9.37430
Epoch: 0232 train_loss= 0.25995 train_acc= 0.88903 val_loss= 0.67166 val_acc= 0.73380 test_acc= 0.73467 time= 9.42031
Epoch: 0233 train_loss= 0.27531 train_acc= 0.88465 val_loss= 0.67780 val_acc= 0.72817 test_acc= 0.73438 time= 9.37990
Epoch: 0234 train_loss= 0.26727 train_acc= 0.88778 val_loss= 0.65669 val_acc= 0.73803 test_acc= 0.73270 time= 9.49047
Epoch: 0235 train_loss= 0.25700 train_acc= 0.89059 val_loss= 0.73397 val_acc= 0.73239 test_acc= 0.72876 time= 9.43395
Epoch: 0236 train_loss= 0.27693 train_acc= 0.87949 val_loss= 0.72715 val_acc= 0.71831 test_acc= 0.72819 time= 9.37171
Epoch: 0237 train_loss= 0.26260 train_acc= 0.88590 val_loss= 0.66580 val_acc= 0.73099 test_acc= 0.73270 time= 9.40740
Epoch: 0238 train_loss= 0.25863 train_acc= 0.88997 val_loss= 0.66178 val_acc= 0.74085 test_acc= 0.74057 time= 9.35542
Epoch: 0239 train_loss= 0.25929 train_acc= 0.89028 val_loss= 0.69102 val_acc= 0.72958 test_acc= 0.73832 time= 9.33036
Epoch: 0240 train_loss= 0.26485 train_acc= 0.88840 val_loss= 0.65980 val_acc= 0.73380 test_acc= 0.73917 time= 9.35281
Epoch: 0241 train_loss= 0.26637 train_acc= 0.88715 val_loss= 0.69797 val_acc= 0.72676 test_acc= 0.73635 time= 9.37982
Epoch: 0242 train_loss= 0.26549 train_acc= 0.88621 val_loss= 0.70432 val_acc= 0.72817 test_acc= 0.74198 time= 9.36356
Epoch: 0243 train_loss= 0.25673 train_acc= 0.89262 val_loss= 0.63438 val_acc= 0.74085 test_acc= 0.74142 time= 9.34668
Epoch: 0244 train_loss= 0.26590 train_acc= 0.88981 val_loss= 0.62542 val_acc= 0.73803 test_acc= 0.73804 time= 9.40314
Epoch: 0245 train_loss= 0.25687 train_acc= 0.88981 val_loss= 0.66705 val_acc= 0.73803 test_acc= 0.74311 time= 9.26673
Epoch: 0246 train_loss= 0.25020 train_acc= 0.89215 val_loss= 0.67204 val_acc= 0.73944 test_acc= 0.73832 time= 9.40202
Epoch: 0247 train_loss= 0.26998 train_acc= 0.88621 val_loss= 0.67130 val_acc= 0.73662 test_acc= 0.74142 time= 9.40909
Epoch: 0248 train_loss= 0.25466 train_acc= 0.88778 val_loss= 0.69664 val_acc= 0.74085 test_acc= 0.74001 time= 9.37671
Epoch: 0249 train_loss= 0.25122 train_acc= 0.89325 val_loss= 0.64389 val_acc= 0.74648 test_acc= 0.73945 time= 9.34132
Epoch: 0250 train_loss= 0.25811 train_acc= 0.89309 val_loss= 0.67433 val_acc= 0.74225 test_acc= 0.74226 time= 9.39195
Epoch: 0251 train_loss= 0.25919 train_acc= 0.89106 val_loss= 0.63643 val_acc= 0.73099 test_acc= 0.73973 time= 9.34966
Epoch: 0252 train_loss= 0.25793 train_acc= 0.89090 val_loss= 0.65691 val_acc= 0.73521 test_acc= 0.74114 time= 9.41477
Epoch: 0253 train_loss= 0.24874 train_acc= 0.89309 val_loss= 0.67438 val_acc= 0.74507 test_acc= 0.73804 time= 9.36380
Epoch: 0254 train_loss= 0.24812 train_acc= 0.89419 val_loss= 0.70219 val_acc= 0.72394 test_acc= 0.74114 time= 9.41626
Epoch: 0255 train_loss= 0.24409 train_acc= 0.89559 val_loss= 0.69852 val_acc= 0.73944 test_acc= 0.73973 time= 9.36738
Epoch: 0256 train_loss= 0.24873 train_acc= 0.89606 val_loss= 0.66105 val_acc= 0.74225 test_acc= 0.73748 time= 9.37715
Epoch: 0257 train_loss= 0.25190 train_acc= 0.89559 val_loss= 0.68010 val_acc= 0.73380 test_acc= 0.73945 time= 9.35769
Epoch: 0258 train_loss= 0.26159 train_acc= 0.88965 val_loss= 0.63750 val_acc= 0.73803 test_acc= 0.74282 time= 9.35684
Epoch: 0259 train_loss= 0.24228 train_acc= 0.89809 val_loss= 0.70676 val_acc= 0.73944 test_acc= 0.73776 time= 9.34766
Epoch: 0260 train_loss= 0.25195 train_acc= 0.89372 val_loss= 0.67741 val_acc= 0.72676 test_acc= 0.74451 time= 9.40812
Epoch: 0261 train_loss= 0.24696 train_acc= 0.89262 val_loss= 0.68972 val_acc= 0.73521 test_acc= 0.74508 time= 9.35904
Epoch: 0262 train_loss= 0.24756 train_acc= 0.89497 val_loss= 0.66983 val_acc= 0.73803 test_acc= 0.74451 time= 9.42902
Epoch: 0263 train_loss= 0.23576 train_acc= 0.90013 val_loss= 0.72569 val_acc= 0.73944 test_acc= 0.74367 time= 9.39227
Epoch: 0264 train_loss= 0.24694 train_acc= 0.89387 val_loss= 0.68797 val_acc= 0.72958 test_acc= 0.74198 time= 9.41038
Epoch: 0265 train_loss= 0.25045 train_acc= 0.89184 val_loss= 0.69657 val_acc= 0.72817 test_acc= 0.73832 time= 9.44044
Epoch: 0266 train_loss= 0.24415 train_acc= 0.89622 val_loss= 0.68202 val_acc= 0.72535 test_acc= 0.74311 time= 9.36992
Epoch: 0267 train_loss= 0.24140 train_acc= 0.89825 val_loss= 0.69179 val_acc= 0.74085 test_acc= 0.74451 time= 9.41314
Epoch: 0268 train_loss= 0.24556 train_acc= 0.89981 val_loss= 0.69714 val_acc= 0.72676 test_acc= 0.74114 time= 9.41774
Epoch: 0269 train_loss= 0.24127 train_acc= 0.89825 val_loss= 0.68328 val_acc= 0.73521 test_acc= 0.74029 time= 9.34645
Epoch: 0270 train_loss= 0.23652 train_acc= 0.89950 val_loss= 0.71008 val_acc= 0.73239 test_acc= 0.74367 time= 9.30617
Epoch: 0271 train_loss= 0.23571 train_acc= 0.90138 val_loss= 0.70447 val_acc= 0.75211 test_acc= 0.74282 time= 9.65783
Epoch: 0272 train_loss= 0.22610 train_acc= 0.90622 val_loss= 0.71326 val_acc= 0.72817 test_acc= 0.74311 time= 9.28041
Epoch: 0273 train_loss= 0.24877 train_acc= 0.89716 val_loss= 0.66448 val_acc= 0.73803 test_acc= 0.74282 time= 9.44393
Epoch: 0274 train_loss= 0.22312 train_acc= 0.90606 val_loss= 0.72372 val_acc= 0.73944 test_acc= 0.74479 time= 9.44585
Epoch: 0275 train_loss= 0.23680 train_acc= 0.90200 val_loss= 0.71687 val_acc= 0.71831 test_acc= 0.73495 time= 9.39549
Epoch: 0276 train_loss= 0.24296 train_acc= 0.89684 val_loss= 0.68692 val_acc= 0.73239 test_acc= 0.74648 time= 9.42836
Epoch: 0277 train_loss= 0.23027 train_acc= 0.90513 val_loss= 0.69398 val_acc= 0.73521 test_acc= 0.74282 time= 9.26210
Epoch: 0278 train_loss= 0.23032 train_acc= 0.90403 val_loss= 0.70078 val_acc= 0.73239 test_acc= 0.74451 time= 9.37999
Epoch: 0279 train_loss= 0.23837 train_acc= 0.90560 val_loss= 0.69338 val_acc= 0.73521 test_acc= 0.74226 time= 9.40872
Epoch: 0280 train_loss= 0.23402 train_acc= 0.90138 val_loss= 0.68410 val_acc= 0.73099 test_acc= 0.74648 time= 9.31526
Epoch: 0281 train_loss= 0.23569 train_acc= 0.90091 val_loss= 0.70239 val_acc= 0.73099 test_acc= 0.73917 time= 9.38258
Epoch: 0282 train_loss= 0.23820 train_acc= 0.90309 val_loss= 0.70577 val_acc= 0.73521 test_acc= 0.74282 time= 9.46656
Epoch: 0283 train_loss= 0.24047 train_acc= 0.89950 val_loss= 0.65842 val_acc= 0.73380 test_acc= 0.74395 time= 9.37246
Epoch: 0284 train_loss= 0.22730 train_acc= 0.90450 val_loss= 0.70678 val_acc= 0.73944 test_acc= 0.73945 time= 9.51553
Epoch: 0285 train_loss= 0.23061 train_acc= 0.90685 val_loss= 0.71630 val_acc= 0.73239 test_acc= 0.74142 time= 9.31654
Epoch: 0286 train_loss= 0.24047 train_acc= 0.89809 val_loss= 0.67403 val_acc= 0.73239 test_acc= 0.74029 time= 9.38103
Epoch: 0287 train_loss= 0.22682 train_acc= 0.91044 val_loss= 0.68074 val_acc= 0.74225 test_acc= 0.74114 time= 9.41290
Epoch: 0288 train_loss= 0.23378 train_acc= 0.90075 val_loss= 0.68912 val_acc= 0.73662 test_acc= 0.74001 time= 9.37079
Epoch: 0289 train_loss= 0.23816 train_acc= 0.90372 val_loss= 0.72509 val_acc= 0.72394 test_acc= 0.73804 time= 9.47469
Epoch: 0290 train_loss= 0.23620 train_acc= 0.89934 val_loss= 0.69028 val_acc= 0.75352 test_acc= 0.73945 time= 9.41381
Epoch: 0291 train_loss= 0.22896 train_acc= 0.89966 val_loss= 0.67896 val_acc= 0.74648 test_acc= 0.74114 time= 9.35215
Epoch: 0292 train_loss= 0.22673 train_acc= 0.90544 val_loss= 0.70703 val_acc= 0.73662 test_acc= 0.74564 time= 9.44908
Epoch: 0293 train_loss= 0.22722 train_acc= 0.90341 val_loss= 0.70001 val_acc= 0.73662 test_acc= 0.73889 time= 9.38804
Epoch: 0294 train_loss= 0.23273 train_acc= 0.90435 val_loss= 0.71646 val_acc= 0.73380 test_acc= 0.74086 time= 9.35751
Epoch: 0295 train_loss= 0.23544 train_acc= 0.90700 val_loss= 0.70393 val_acc= 0.74930 test_acc= 0.74339 time= 9.45055
Epoch: 0296 train_loss= 0.22091 train_acc= 0.90763 val_loss= 0.68733 val_acc= 0.74648 test_acc= 0.74226 time= 9.33286
Epoch: 0297 train_loss= 0.22432 train_acc= 0.90935 val_loss= 0.70470 val_acc= 0.75352 test_acc= 0.74282 time= 9.45516
Epoch: 0298 train_loss= 0.22138 train_acc= 0.90638 val_loss= 0.69942 val_acc= 0.73944 test_acc= 0.74198 time= 9.38376
Epoch: 0299 train_loss= 0.22233 train_acc= 0.90982 val_loss= 0.72312 val_acc= 0.73521 test_acc= 0.74086 time= 9.50871
Optimization Finished!
Best epoch: 297
Test set results: cost= 0.78937 accuracy= 0.74282
Test Precision, Recall and F1-Score...
precision recall f1-score support

       0     0.7372    0.7546    0.7458      1777
       1     0.7487    0.7310    0.7397      1777

accuracy                         0.7428      3554

macro avg 0.7430 0.7428 0.7428 3554
weighted avg 0.7430 0.7428 0.7428 3554

Macro average Test Precision, Recall and F1-Score...
(0.7429607109077572, 0.7428249859313449, 0.7427890645389335, None)
Micro average Test Precision, Recall and F1-Score...
(0.7428249859313449, 0.7428249859313449, 0.742824985931345, None)

[Magicat128]

@whateverud
您好，看起来好像计算精度有些问题，请问有在其他机器上运行过吗？比如linux系统

[susanhhhhhh]

你好，我看这个val的loss是不是过拟合了呀，不知道该怎么解决