https://github.com/ramnathaniel/models-files-manager
Models Files Manager helps you manage saved ML models on your file system, encoding fields into file names
https://github.com/ramnathaniel/models-files-manager
deep-learning keras models pytorch tensorflow
Last synced: 3 months ago
JSON representation
Models Files Manager helps you manage saved ML models on your file system, encoding fields into file names
- Host: GitHub
- URL: https://github.com/ramnathaniel/models-files-manager
- Owner: RamNathaniel
- License: mit
- Created: 2023-04-30T19:43:55.000Z (about 3 years ago)
- Default Branch: master
- Last Pushed: 2023-09-05T20:48:44.000Z (almost 3 years ago)
- Last Synced: 2025-02-16T20:03:23.722Z (over 1 year ago)
- Topics: deep-learning, keras, models, pytorch, tensorflow
- Language: Python
- Homepage:
- Size: 8.79 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Models Files Manager
The Models Files Manager is a simple tool that helps you keep track of saved models (PyTorch/Keras/Onnx/etc.) saved on the hard drive. You can set up a number of fields, that will get encoded into the file names, and query the files saved.
The Models Files Manager helps you avoid illegal chars in the paths, and easily retrive saved files for loading.
## Installation
You can install the Models Files Manager from [PyPI](https://pypi.org/project/models-files-manager/):
python -m pip install models-files-manager
The manager is supported and tested on Python 3.8 and above.
## How to use
To use the Models Files Manager, in your own Python code, by importing from the `models_files_manager` package:
>>> from models_files_manager import ModelsFilesManager
>>> manager = ModelsFilesManager(root, ['field1', 'field2'], 'ext')
>>> models_file_names = manager.get_model_files(lambda d: d['field1'] == 'value1')
Or to create a new file name:
>>> from models_files_manager import ModelsFilesManager
>>> manager = ModelsFilesManager(root, ['field1', 'field2'], 'ext')
>>> file_name = manager.get_file_name({'field1': 'value1', 'field2': 'value2'})
Or get the field values of a specific model, from the path:
>>> from models_files_manager import ModelsFilesManager
>>> manager = ModelsFilesManager(root, ['field1', 'field2'], 'ext')
>>> fields = manager.get_fields_values('model file name')
>>> print(fields['field1'])
To directly encode/decode into strings to a path-safe format:
>>> from models_files_manager import ModelsFilesManager
>>> safe_string = ModelsFilesManager.path_escape(string)
>>> orig_string = ModelsFilesManager.path_decode(safe_string)
Free to use in any way, open source or commercial - see attached license.
Happy training!
Ram Nathaniel, 2023