Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/toodef/neural-pipeline
Neural networks training pipeline based on PyTorch
https://github.com/toodef/neural-pipeline
deep-learning image-classification image-segmentation neural-networks object-detection pipeline pytorch supervised-learning training-pipeline
Last synced: 4 months ago
JSON representation
Neural networks training pipeline based on PyTorch
- Host: GitHub
- URL: https://github.com/toodef/neural-pipeline
- Owner: toodef
- License: mit
- Archived: true
- Created: 2018-12-14T20:24:38.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2020-06-01T17:50:58.000Z (over 4 years ago)
- Last Synced: 2024-10-14T06:38:21.179Z (4 months ago)
- Topics: deep-learning, image-classification, image-segmentation, neural-networks, object-detection, pipeline, pytorch, supervised-learning, training-pipeline
- Language: Python
- Homepage: https://neural-pipeline.readthedocs.io
- Size: 13.6 MB
- Stars: 312
- Watchers: 7
- Forks: 24
- Open Issues: 29
-
Metadata Files:
- Readme: README.md
- Contributing: docs/CONTRIBUTING.rst
- License: LICENSE
Awesome Lists containing this project
README
# This project are closed. New versions and much more are there: [PiePline](https://github.com/PiePline)
# Neural Pipeline
Neural networks training pipeline based on PyTorch. Designed to standardize training process and accelerate experiments.
[](https://travis-ci.org/toodef/neural-pipeline)
[](https://coveralls.io/github/toodef/neural-pipeline?branch=master)
[](https://codeclimate.com/github/toodef/neural-pipeline/maintainability)
[](https://gitter.im/neural-pipeline/community)* Core is about 2K lines, covered by tests, that you don't need to write again
* Flexible and customizable training process
* Checkpoints management and train process resuming (source and target device independent)
* Metrics processing and visualization by builtin ([tensorboard](https://www.tensorflow.org/guide/summaries_and_tensorboard), [Matplotlib](https://matplotlib.org)) or custom monitors
* Training best practices (e.g. learning rate decaying and hard negative mining)
* Metrics logging and comparison (DVC compatible)# Getting started:
### Documentation
[](https://neural-pipeline.readthedocs.io/en/master/?badge=master)
* [See the full documentation there](https://neural-pipeline.readthedocs.io/en/master/)
* [Read getting started guide](https://neural-pipeline.readthedocs.io/en/master/getting_started/index.html)### See the examples
* MNIST classification - [notebook](https://github.com/toodef/neural-pipeline/blob/master/examples/notebooks/img_classification.ipynb), [file](https://github.com/toodef/neural-pipeline/blob/master/examples/files/img_classification.py), [Kaggle kernel](https://www.kaggle.com/toodef/cnn-training-with-less-code)
* Segmentation - [notebook](https://github.com/toodef/neural-pipeline/blob/master/examples/notebooks/img_segmentation.ipynb), [file](https://github.com/toodef/neural-pipeline/blob/master/examples/files/img_segmentation.py)
* Resume training process - [file](https://github.com/toodef/neural-pipeline/blob/master/examples/files/resume_train.py)### Neural Pipeline short overview:
```python
import torchfrom neural_pipeline.builtin.monitors.tensorboard import TensorboardMonitor
from neural_pipeline.monitoring import LogMonitor
from neural_pipeline import DataProducer, TrainConfig, TrainStage,\
ValidationStage, Trainer, FileStructManagerfrom somethig import MyNet, MyDataset
fsm = FileStructManager(base_dir='data', is_continue=False)
model = MyNet().cuda()train_dataset = DataProducer([MyDataset()], batch_size=4, num_workers=2)
validation_dataset = DataProducer([MyDataset()], batch_size=4, num_workers=2)train_config = TrainConfig(model, [TrainStage(train_dataset),
ValidationStage(validation_dataset)], torch.nn.NLLLoss(),
torch.optim.SGD(model.parameters(), lr=1e-4, momentum=0.5))trainer = Trainer(train_config, fsm, torch.device('cuda:0')).set_epoch_num(50)
trainer.monitor_hub.add_monitor(TensorboardMonitor(fsm, is_continue=False))\
.add_monitor(LogMonitor(fsm))
trainer.train()
```
This example of training MyNet on MyDataset with vizualisation in Tensorflow and with metrics logging for further experiments comparison.# Installation:
[](https://badge.fury.io/py/neural-pipeline)
[](https://pepy.tech/project/neural-pipeline)
[](https://pepy.tech/project/neural-pipeline)`pip install neural-pipeline`
##### For `builtin` module using install:
`pip install tensorboardX matplotlib`##### Install latest version before it's published on PyPi
`pip install -U git+https://github.com/toodef/neural-pipeline`