brightmart / ai_law Goto Github PK

Can your provide the trained checkpoint to reconstruct your result?

请问在large数据集上是怎么划分train/test/valid的呢

great job!
but when I use your code to train the 'big_data' (1700k),cpu always occupied 100%,then I cannot connect to my server.

您好，最近参加了一个长文本分类比赛，文本的平均长度为720个词，请问有什么模型可以推荐吗？开始用的DPCNN来做，由于内存才10g左右只能用很小的batch来做，并且很慢。多谢

tensorflow.python.framework.errors_impl.DataLossError: Checksum does not match: stored 2982177973 vs. calculated on the restored bytes 1718988051
我上传之后总是这个问题，请问您有什么见解。
我在本地测试都不会出现问题，但是就出在上传会报这个错误。
为了验证是我代码的问题，还是上传出错。我今天把您的代码跑了一篇，然后在本地跑main.py正常，然后压缩上传了一次，同样出现了这种错误。我压缩的方式是zip -r -9 predictor.zip predictor/,我想既然我上传您的代码也会出现同样的错误，那错误应该不是出在代码里，而是压缩的步骤或上传的步骤中存在问题，请问您是怎么压缩文件上传的？

train:
tf.app.flags.DEFINE_string("ckpt_dir","./predictor/checkpoint/","checkpoint location for the model")
tf.app.flags.DEFINE_string("model","text_cnn","name of model:han,text_cnn,dp_cnn,c_gru,c_gru2,gru,pooling")
predictor:
tf.app.flags.DEFINE_string("model_dpcnn", "dp_cnn", "name of model:han,c_gru,c_gru2,gru,text_cnn")
tf.app.flags.DEFINE_string("ckpt_dir_dpcnn", "predictor/checkpoint_dpcnn_big32/", "checkpoint location for the model")

tf.app.flags.DEFINE_string("model", "text_cnn", "name of model:han,c_gru,c_gru2,gru,text_cnn")
tf.app.flags.DEFINE_string("ckpt_dir", "predictor/checkpoint/", "checkpoint location for the model")

当训练模型的时候是用text_cnn, 不理解在预测的时候为什么是用的text_cnn和dpcnn. 难道是在训练模型的时候，需要再训练dpcnn吗？谢谢您的帮助。

('using pre-trained word emebedding.started.word2vec_model_path:', 'data/news_12g_baidubaike_20g_novel_90g_embedding_64.bin')
Traceback (most recent call last):
File "/home/qiu/PycharmProjects/ai_law-master (2)/HAN_train.py", line 271, in
tf.app.run()
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/platform/app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "/home/qiu/PycharmProjects/ai_law-master (2)/HAN_train.py", line 101, in main
assign_pretrained_word_embedding(sess, vocabulary_index2word, vocab_size, model,FLAGS.word2vec_model_path,model.Embedding)
File "/home/qiu/PycharmProjects/ai_law-master (2)/HAN_train.py", line 238, in assign_pretrained_word_embedding
for word, vector in zip(word2vec_model.vocab, word2vec_model.vectors):
AttributeError: 'EuclideanKeyedVectors' object has no attribute 'vectors'

您好，请问Imbalance Classification for Skew Data中的threshold是如何确定的。为什么是num_mini_example/freq