rjt1990 / mantra Goto Github PK

File "/home/ubuntu/anaconda3/envs/tensorflow_p36/lib/python3.6/site-packages/mantraml/data/Dataset.py", line 288, in extract_file_data
file_hashes = self.get_data_dependency_hashes(is_extract_folder=is_extract_folder, is_hash=is_hash)
File "/home/ubuntu/anaconda3/envs/tensorflow_p36/lib/python3.6/site-packages/mantraml/data/Dataset.py", line 401, in get_data_dependency_hashes
self.extract_tar_file(file_path)
File "/home/ubuntu/anaconda3/envs/tensorflow_p36/lib/python3.6/site-packages/mantraml/data/Dataset.py", line 349, in extract_tar_file
os.rename('%sraw/%s' % (self.data_dir, top_level_dir), '%sraw/%s' % (self.data_dir, '.extract'))
FileNotFoundError: [Errno 2] No such file or directory: '/home/ubuntu/finalproject/data/horse2zebra/raw/t' -> '/home/ubuntu/finalproject/data/horse2zebra/raw/.extract'

Is supporting scikit-learn models (even if they are usually not Deep Learning models) on the roadmap? If so, what are the minimum requirements to be implemented for support?

Errors out with an OverFlow file when the dataset dependencies are large files

This works if training is run again - but not on the first upload.

Currently does not sort dates correctly - string based sorting; should be modified to work with dates (or timestamp ints).

flake8 testing of https://github.com/RJT1990/mantra on Python 3.7.0

$ flake8 . --count --select=E901,E999,F821,F822,F823 --show-source --statistics

/home/travis/virtualenv/python3.7.0/lib/python3.7/site-packages/pycodestyle.py:113: FutureWarning: Possible nested set at position 1
  EXTRANEOUS_WHITESPACE_REGEX = re.compile(r'[[({] | []}),;:]')
./mantraml/templates/models/pytorch/base/model.py:14:21: F821 undefined name 'dask'
        self.data = dask
                    ^
./mantraml/templates/models/keras/base/model.py:14:21: F821 undefined name 'dask'
        self.data = dask
                    ^
./mantraml/templates/models/tensorflow/base/model.py:14:21: F821 undefined name 'dask'
        self.data = dask
                    ^
./mantraml/templates/models/default/base/model.py:15:21: F821 undefined name 'dask'
        self.data = dask
                    ^
./mantraml/tests/test_model_tensorflow.py:14:21: F821 undefined name 'dask'
        self.data = dask
                    ^
./mantraml/ui/core/views.py:358:48: F821 undefined name 'trial_contents'
            new_contents = [trial for trial in trial_contents if not trial[2] == trial_hash]
                                               ^
./mantraml/ui/core/views.py:361:57: F821 undefined name 'trial_contents'
                trial_folder_name = [trial for trial in trial_contents if trial[2] == trial_hash][0][1]
                                                        ^
7     F821 undefined name 'dask'
7

Defaults to 4 colour dimensions - can also be 3 in some instances.

Additionally, in the ImageTemplate there is a class variable called normalized - should be normalize

Currently the user must specify --dev in addition to --cloud to do training on a development instance - an instance that does not close after training.

Since the development machine is where most interactions with cloud environments takes place, it makes more sense to make this the default option, and for the user to only specify temporary instances when they are doing a job that they want to run to completion (once they are satisfied with the model).

The project looks quite interesting - do you have a roadmap that you are willing to share on what you want to achieve in the next 6/12 months?

It would probably help people to contribute to the project itself :)

The requirements should also include the following:

tensorflow
h5py