Skip to content

local variable 't_words' referenced before assignment #2

@HaozhengLi

Description

@HaozhengLi

I use the Twitter_DataSet to train.

I successfully train a df_model but get errors while training a rl_model.

2019-06-19 21:55:53 Loading data ...
2 data loaded
2019-06-19 21:56:52 Data loaded.
2019-06-19 21:56:52.885374: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
300 100 100 100 2 2
df_saved\model is restored.
in RL the begining
2019-06-19 21:56:53 Now Start RL training ...
0 0
Traceback (most recent call last):
  File "main.py", line 165, in <module>
    rl_train(sess, mm, 0.5, 50000)
  File "main.py", line 84, in rl_train
    x, y, ids, seq_states, max_id = get_rl_batch(ids, seq_states, isStop, max_id, 0, 3150)
  File "E:\Output\Python_output\ERD\ERD\dataUtils.py", line 187, in get_rl_batch
    try:
UnboundLocalError: local variable 't_words' referenced before assignment

Here's my question, in <dataUtils.py> line 181 we try to get <t_words>.
But if we fail, there will not be a reachable <t_words>, then errors occur.

        else:
            try:
                t_words = data[ids[i]]['text'][seq_states[i]].strip().split(" ")
            except:
                print(ids[i],seq_states[i])
            for j in range(len(t_words)):
                m_word = t_words[j]
                try:
                    input_x[i][j] = word2vec[m_word]
                except:
                    miss_vec = 1
            input_y[i] = data_y[ids[i]]

So how can I run the code correctly?

Thank you.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions