TypeError: repl must be a string

[LuciferA678]

As I started training Deepspeech Model it gave error
Traceback (most recent call last):
File "./DeepSpeech.py", line 934, in
tf.app.run(main)
File "/root/anaconda3/lib/python3.6/site-packages/tensorflow/python/platform/app.py", line 125, in run
_sys.exit(main(argv))
File "./DeepSpeech.py", line 886, in main
train()
File "./DeepSpeech.py", line 386, in train
hdf5_cache_path=FLAGS.train_cached_features_path)
File "/app/DeepSpeech-master/util/preprocess.py", line 58, in preprocess
file['wav_filename'] = file['wav_filename'].str.replace(r'(^[^/])', lambda m: os.path.join(csv_dir, m.group(1)))
File "/root/anaconda3/lib/python3.6/site-packages/pandas/core/strings.py", line 1506, in replace
flags=flags)
File "/root/anaconda3/lib/python3.6/site-packages/pandas/core/strings.py", line 322, in str_replace
raise TypeError("repl must be a string")
TypeError: repl must be a string
TypeError: repl must be a string
what should be done?

[kdavis-mozilla]

What command line did you execute?

[LuciferA678]

./DeepSpeech.py --train_files /app/cv-valid-train.csv --dev_files /app/cv-valid-dev.csv --test_files /app/cv-valid-test.csv

[kdavis-mozilla]

How are cv-valid-train.csv, cv-valid-dev.csv, and cv-valid-test.csv created?

[LuciferA678]

cv-valid-train.csv, cv-valid-dev.csv, and cv-valid-test.csv are created by bin/import_cv.py script

[kdavis-mozilla]

Running on what data?

[LuciferA678]

data which was downloaded by bin/import_cv.py ,i.e cv_corpus_v1

[kdavis-mozilla]

What version of pandas are you using?

[LuciferA678]

pandas =0.19.2

[kdavis-mozilla]

That's the problem.

This bug was fixed two years ago in Pandas 0.20.0 , see this issue. Could you update to a recent, Pandas.

[lissyx]

That's the problem.

This bug was fixed two years ago in Pandas 0.20.0 , see this issue. Could you update to a recent, Pandas.

Maybe we should enforce pandas >= 0.20.0 in requirements.txt ?

[kdavis-mozilla]

Maybe, but I don't think it would have fixed this problem. LuciferA678 didn't use requirements.txt but used an anaconda package with a pre-installed pandas.

[LuciferA678]

Thanks a lot @kdavis-mozilla it worked,and I did not used requirements.txt
but I landed in another error
Traceback (most recent call last):
File "./DeepSpeech.py", line 934, in
tf.app.run(main)
File "/root/anaconda3/lib/python3.6/site-packages/tensorflow/python/platform/app.py", line 125, in run
_sys.exit(main(argv))
File "./DeepSpeech.py", line 886, in main
train()
File "./DeepSpeech.py", line 424, in train
gradients, loss = get_tower_results(model_feeder, optimizer, dropout_rates)
File "./DeepSpeech.py", line 256, in get_tower_results
avg_loss = calculate_mean_edit_distance_and_loss(model_feeder, i, dropout_rates, reuse=i>0)
File "./DeepSpeech.py", line 186, in calculate_mean_edit_distance_and_loss
logits, _ = BiRNN(batch_x, batch_seq_len, dropout, reuse)
File "./DeepSpeech.py", line 88, in BiRNN
layer_1 = tf.nn.dropout(layer_1, rate=dropout[0])
TypeError: dropout() got an unexpected keyword argument 'rate'
Can You help it out?

[kdavis-mozilla]

Could you use requirements.txt as indicted in Installing prerequisites for training.

[LuciferA678]

Thank you the issue is solved and trainning is started but
after 55 epochs error occured

:---->E [Errno 5] Input/output error

What is this error and what should be done?