I was trying to reproduce your experiments but I am having trouble finding the embeddings location.
I changed the argument parameters for train.py from --train --val --test parameters to --ner_train_file, --ner_dev_file etc. (Also, docker build crashes at pip install reqirements dynet specification, worked after removing the version for dynet.)
this is the error I am getting for nor specifying the location of pretrained embeddings
File "/joint-ner-and-md-tagger/utils/train.py", line 57, in train
data_dict, id_to_tag, word_to_id, stats_dict = prepare_datasets(model, opts, parameters)
File "/joint-ner-and-md-tagger/utils/loader.py", line 842, in prepare_datasets
_prepare_datasets(opts, parameters, for_training=for_training)
File "/joint-ner-and-md-tagger/utils/loader.py", line 754, in _prepare_datasets
load_sentences(opts_dict[label+"_train_file"], zeros, parameters['file_format'])
File "/joint-ner-and-md-tagger/utils/loader.py", line 33, in load_sentences
input_f = codecs.open(input_file_path_or_list, 'r', 'utf8')
File "miniconda3/envs/boun/lib/python2.7/codecs.py", line 898, in open
file = __builtin__.open(filename, mode, buffering)
IOError: [Errno 2] No such file or directory: ''