search

koth / kcws

Deep Learning Chinese Word Segment
2.08k stars 648 forks source link

IndexError: index 0 is out of bounds for axis 0 with size 0 train_cws_lstm.py 中报的错，一直没有解决 #57

Open LubbyAnneLiu opened 7 years ago

LubbyAnneLiu commented 7 years ago

[100] loss: [28.518539] [200] loss: [17.184589] [300] loss: [14.19612] [400] loss: [13.810515] [500] loss: [11.092119] [600] loss: [10.731837] [700] loss: [10.634817] [800] loss: [8.1204176] [900] loss: [8.273036] [1000] loss: [8.4017849] Traceback (most recent call last): File "train_cws_lstm_PKU0421.py", line 288, in tf.app.run() File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/platform/app.py", line 43, in run sys.exit(main(sys.argv[:1] + flags_passthrough)) File "train_cws_lstm_PKU0421.py", line 278, in main model.inp, tX, tY) File "train_cws_lstm_PKU0421.py", line 224, in test_evaluate tf_unaryscores, transMatrix) File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/crf/python/ops/crf.py", line 298, in viterbi_decode trellis[0] = score[0] IndexError: index 0 is out of bounds for axis 0 with size 0

大神帮帮忙 每次走到test_evaluate()都会报这个错，我看bug出现在crf.py里，实在是不懂，求指点

koth commented 7 years ago

应该不是bug，可能那个地方没正确传入参数，比如tag数之类的

LubbyAnneLiu commented 7 years ago

对对对 我知道不是模型的问题 大概就是传参问题 我试着输出了这一步的传入viterbi_decode（）函数的参数tf_unary_scores_和 tf_unary_scores_和transMatrix，好像也没啥问题呀。下面是传入的参数。还有可能是什么问题呀？实在是搞不明白~~拜托~ [[ 5.10233212 -0.54984224 -3.3073473 -2.23302674] [ 5.00223017 1.7417382 -4.27542925 -3.73243022] [ 0.62986803 3.85293174 -1.22212958 -3.96823788] [ 0.70197666 -2.35121202 -0.62936413 2.68005323] [ 2.77736568 -0.27267972 -2.36598563 -0.48937809] [ 0.8004145 2.19753051 -2.04182363 -1.4268291 ] [ 1.86268508 -2.15325522 -1.77105057 2.32386279] [ 4.85579681 -1.55591941 -2.76405716 -1.15710473]] transMatrix [[ 0.19105484 0.18477149 -0.04392412 0.21299016] [-0.58652729 -0.39606482 0.41640422 0.20975219] [ 0.09511784 0.22878556 0.03559215 0.52470607] [ 0.66774857 0.75008577 -0.61195064 0.34702992]] tf_unaryscores [[ 1.15849197 2.9072063 -1.57891667 -3.05160451] [ 1.61692488 -2.64840722 -1.64531779 2.63486719] [ 5.21047401 -1.04094565 -2.8609128 -2.224334 ] [ 2.42167044 3.55341673 -2.2693305 -4.78759813] [-1.05136931 -0.18433641 1.11184239 0.26811025] [ 1.70061576 -1.69640815 -1.63032782 1.7006048 ] [ 4.50150728 -2.17166281 -2.26328468 -0.56402373]] transMatrix [[ 0.19105484 0.18477149 -0.04392412 0.21299016] [-0.58652729 -0.39606482 0.41640422 0.20975219] [ 0.09511784 0.22878556 0.03559215 0.52470607] [ 0.66774857 0.75008577 -0.61195064 0.34702992]]