зеркало из https://github.com/microsoft/MT-DNN.git
9dec8253f1
* MT-DNN Feature release v1.0.0 ignore: IDE meta files add: initial checkin feat: make package pip installable doc: add contribution steps and update readme with additional information fix: wrong file references add: package dependencies like PyTorch transformer support fix import statements feat: add download utils feat: add download shell script doc: update readme with testing instructions doc: add data downloading and processing step feat: move fit and predict into MTDNN Model feat: make logger create a new log file each run remove: fit and predict functions from the pipeline class update: remove stale references doc: fit and predict now on the model object formatting code snippet doc: add pip install steps ignore: checkpoints and log files add: conda file for env generation feat: docker file support doc: README for example and data add: sample data in json lines format feat: jupyter example remove: batch_size is controlled by configuration feat: jupyter notebook add: license and copyright add: git lfs track sample data files data: add sample data file with git lfs doc: batch_size is now set in config cleanup |
||
|---|---|---|
| .. | ||
| common | ||
| tasks | ||
| __init__.py | ||
| configuration_mtdnn.py | ||
| dataset_mtdnn.py | ||
| modeling_mtdnn.py | ||
| process_mtdnn.py | ||