I am trying to create an Arabic audio model using about 80,000 audio files.
This is the error I am having while trying to train. Can you help me know if something is not right with the parameters I’m assigning or something else? :
python -u DeepSpeech.py --train_files /Users/naveen/Downloads/all_datasets/DeepSpeech/TRAIN/train.csv --dev_files /Users/naveen/Downloads/all_datasets/DeepSpeech/DEV/dev.csv --test_files /Users/naveen/Downloads/all_datasets/DeepSpeech/TEST/test.csv --train_batch_size 80 --dev_batch_size 80 --test_batch_size 40 --n_hidden 375 --epoch 33 --validation_step 1 --early_stop True --earlystop_nsteps 6 --estop_mean_thresh 0.1 --estop_std_thresh 0.1 --dropout_rate 0.22 --learning_rate 0.00095 --report_count 100 --use_seq_length False --export_dir /Users/naveen/Downloads/all_datasets/DeepSpeech/results/model_export/ --checkpoint_dir /Users/naveen/Downloads/all_datasets/DeepSpeech/results/checkout/ --decoder_library_path /Users/naveen/Downloads/DeepSpeech/DeepSpeech/libctc_decoder_with_kenlm.so --alphabet_config_path /Users/naveen/Downloads/all_datasets/DeepSpeech/alphabet.txt --lm_binary_path /Users/naveen/Downloads/all_datasets/DeepSpeech/lm.binary --lm_trie_path /Users/naveen/Downloads/all_datasets/DeepSpeech/trie
/Users/naveen/anaconda3/lib/python3.6/site-packages/h5py/init.py:36: FutureWarning: Conversion of the second argument of issubdtype from float to np.floating is deprecated. In future, it will be treated as np.float64 == np.dtype(float).type.
from ._conv import register_converters as _register_converters
Traceback (most recent call last):
File “/Users/naveen/Downloads/DeepSpeech/DeepSpeech/util/audio.py”, line 7, in
from deepspeech.utils import audioToInputVector
ModuleNotFoundError: No module named ‘deepspeech’
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File “DeepSpeech.py”, line 24, in
from util.audio import audiofile_to_input_vector
File “/Users/naveen/Downloads/DeepSpeech/DeepSpeech/util/audio.py”, line 10, in
from python_speech_features import mfcc
ModuleNotFoundError: No module named ‘python_speech_features’
Deepak’s MacBook Pro:DeepSpeech naveen$ pip install deepspeech
Collecting deepspeech
Using cached https://files.pythonhosted.org/packages/14/c9/e969fbdaac6b2ce7a0fc4c24f0bc96ab4aaaac0e5c0be85f0dceb90c6fb9/deepspeech-0.1.1-cp36-cp36m-macosx_10_10_x86_64.whl
Requirement already satisfied: scipy in /Users/naveen/anaconda3/lib/python3.6/site-packages (from deepspeech)
Requirement already satisfied: numpy in /Users/naveen/anaconda3/lib/python3.6/site-packages (from deepspeech)
Installing collected packages: deepspeech
Successfully installed deepspeech-0.1.1
You are using pip version 9.0.1, however version 10.0.1 is available.
You should consider upgrading via the ‘pip install --upgrade pip’ command.
Deepak’s MacBook Pro:DeepSpeech naveen$ sh run_file.sh