mahmoodn opened a new issue #14096: module 'mxnet.symbol' has no attribute 'WarpCTC' URL: https://github.com/apache/incubator-mxnet/issues/14096 I want to run speech recognition example and I have followed the steps in the readme file. Here are the commands I ran to prepare the example ``` pip install mxboard pip install soundfile git clone https://github.com/baidu-research/warp-ctc.git cd warp-ctc/ && mkdir build && cd build/ cmake .. && make ``` So, I can verify the preparations ``` $ cat Libri_sample.json {"duration": 2.9450625, "text": "and sharing her house which was near by", "key": "./Libri_sample/3830-12531-0030.wav"} {"duration": 3.94, "text": "we were able to impart the information that we wanted", "key": "./Libri_sample/3830-12529-0005.wav"} $ ls Libri_sample 3830-12529-0005.wav 3830-12531-0030.wav $ echo $LD_LIBRARY_PATH /home/mahmood/mx/mxnet/example/speech_recognition/warp-ctc/build::/usr/local/cuda-10.0/lib64 ``` However, I the run isn't successful and I get the following error ``` $ python main.py --configfile default.cfg ================================================================================ [ DEBUG][2019/02/08 16:55:28.242] LabelUtil init [ INFO][2019/02/08 16:55:28.243] Reading description file: ./Libri_sample.json for partition: train [ INFO][2019/02/08 16:55:28.243] Reading description file: ./Libri_sample.json for partition: validation [ INFO][2019/02/08 16:55:28.243] Generate mean and std from samples [ INFO][2019/02/08 16:55:28.243] Calculating mean and std from samples [ INFO][2019/02/08 16:55:28.356] End calculating mean and std from samples [ INFO][2019/02/08 16:55:28.372] Config: [ INFO][2019/02/08 16:55:28.373] [common] [ INFO][2019/02/08 16:55:28.373] mode = train [ INFO][2019/02/08 16:55:28.373] context = gpu0 [ INFO][2019/02/08 16:55:28.373] prefix = test_fc [ INFO][2019/02/08 16:55:28.373] model_file = test_fc-0040 [ INFO][2019/02/08 16:55:28.373] batch_size = 2 [ INFO][2019/02/08 16:55:28.373] log_filename = test.log [ INFO][2019/02/08 16:55:28.373] save_checkpoint_every_n_epoch = 20 [ INFO][2019/02/08 16:55:28.373] save_checkpoint_every_n_batch = 1000 [ INFO][2019/02/08 16:55:28.373] is_bi_graphemes = False [ INFO][2019/02/08 16:55:28.373] mxboard_log_dir = mxlog/libri_sample [ INFO][2019/02/08 16:55:28.373] mx_random_seed = 1234 [ INFO][2019/02/08 16:55:28.373] random_seed = 1234 [ INFO][2019/02/08 16:55:28.373] kvstore_option = device [ INFO][2019/02/08 16:55:28.373] [ INFO][2019/02/08 16:55:28.374] [data] [ INFO][2019/02/08 16:55:28.374] max_duration = 16.0 [ INFO][2019/02/08 16:55:28.374] train_json = ./Libri_sample.json [ INFO][2019/02/08 16:55:28.374] test_json = ./Libri_sample.json [ INFO][2019/02/08 16:55:28.374] val_json = ./Libri_sample.json [ INFO][2019/02/08 16:55:28.374] language = en [ INFO][2019/02/08 16:55:28.374] width = 161 [ INFO][2019/02/08 16:55:28.374] height = 1 [ INFO][2019/02/08 16:55:28.374] channel = 1 [ INFO][2019/02/08 16:55:28.374] stride = 1 [ INFO][2019/02/08 16:55:28.374] [ INFO][2019/02/08 16:55:28.374] [arch] [ INFO][2019/02/08 16:55:28.374] channel_num = 32 [ INFO][2019/02/08 16:55:28.374] conv_layer1_filter_dim = [11, 41] [ INFO][2019/02/08 16:55:28.374] conv_layer1_stride = [2, 2] [ INFO][2019/02/08 16:55:28.374] conv_layer2_filter_dim = [11, 21] [ INFO][2019/02/08 16:55:28.374] conv_layer2_stride = [1, 2] [ INFO][2019/02/08 16:55:28.375] num_rnn_layer = 1 [ INFO][2019/02/08 16:55:28.375] num_hidden_rnn_list = [1760] [ INFO][2019/02/08 16:55:28.375] num_hidden_proj = 0 [ INFO][2019/02/08 16:55:28.375] num_rear_fc_layers = 0 [ INFO][2019/02/08 16:55:28.375] num_hidden_rear_fc_list = [] [ INFO][2019/02/08 16:55:28.375] act_type_rear_fc_list = [] [ INFO][2019/02/08 16:55:28.375] rnn_type = bigru [ INFO][2019/02/08 16:55:28.375] lstm_type = fc_lstm [ INFO][2019/02/08 16:55:28.375] is_batchnorm = True [ INFO][2019/02/08 16:55:28.375] is_bucketing = False [ INFO][2019/02/08 16:55:28.375] buckets = [] [ INFO][2019/02/08 16:55:28.375] arch_file = arch_deepspeech [ INFO][2019/02/08 16:55:28.375] n_classes = 28 [ INFO][2019/02/08 16:55:28.375] max_t_count = 393 [ INFO][2019/02/08 16:55:28.375] max_label_length = 53 [ INFO][2019/02/08 16:55:28.375] [ INFO][2019/02/08 16:55:28.375] [train] [ INFO][2019/02/08 16:55:28.375] num_epoch = 50 [ INFO][2019/02/08 16:55:28.376] learning_rate = 0.005 [ INFO][2019/02/08 16:55:28.376] learning_rate_annealing = 1.1 [ INFO][2019/02/08 16:55:28.376] initializer = Xavier [ INFO][2019/02/08 16:55:28.376] init_scale = 2 [ INFO][2019/02/08 16:55:28.376] factor_type = in [ INFO][2019/02/08 16:55:28.376] show_every = 1 [ INFO][2019/02/08 16:55:28.376] save_optimizer_states = True [ INFO][2019/02/08 16:55:28.376] normalize_target_k = 2 [ INFO][2019/02/08 16:55:28.376] overwrite_meta_files = True [ INFO][2019/02/08 16:55:28.376] overwrite_bi_graphemes_dictionary = False [ INFO][2019/02/08 16:55:28.376] save_feature_as_csvfile = False [ INFO][2019/02/08 16:55:28.376] enable_logging_train_metric = True [ INFO][2019/02/08 16:55:28.376] enable_logging_validation_metric = True [ INFO][2019/02/08 16:55:28.376] [ INFO][2019/02/08 16:55:28.376] [load] [ INFO][2019/02/08 16:55:28.376] load_optimizer_states = True [ INFO][2019/02/08 16:55:28.376] is_start_from_batch = False [ INFO][2019/02/08 16:55:28.376] [ INFO][2019/02/08 16:55:28.377] [optimizer] [ INFO][2019/02/08 16:55:28.377] optimizer = adam [ INFO][2019/02/08 16:55:28.377] optimizer_params_dictionary = {"beta1":0.9,"beta2":0.999} [ INFO][2019/02/08 16:55:28.377] clip_gradient = 0 [ INFO][2019/02/08 16:55:28.377] weight_decay = 0. [ INFO][2019/02/08 16:55:28.377] Traceback (most recent call last): File "main.py", line 310, in <module> model_loaded, model_num_epoch = load_model(args, contexts, data_train) File "main.py", line 236, in load_model model_loaded = symbol_template.arch(args) File "/home/mahmood/mx/mxnet/example/speech_recognition/arch_deepspeech.py", line 186, in arch (args.config.getint('arch', 'n_classes') + 1)) File "/home/mahmood/mx/mxnet/example/speech_recognition/stt_layer_warpctc.py", line 37, in warpctc_layer net = mx.sym.WarpCTC(data=net, label=label, label_length=num_label, input_length=seq_len) AttributeError: module 'mxnet.symbol' has no attribute 'WarpCTC' ``` What is missing then?
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
