Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/rvk007/deepnet

A python library for computer vision applications
https://github.com/rvk007/deepnet

bce-dice bce-rmse cifar10 cnn computer-vision deep-learning gradcam lrfinder mask-detection monocular-depth-estimation object-recognition python pytorch

Last synced: 11 months ago
JSON representation

A python library for computer vision applications

Awesome Lists containing this project

README 

          


  tensornet

  





[![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1g63kM2rq3pktpTx5neqNlbSVYeT9xvEk)  



Deepnet is an open-source library that can be used for solving problems of Computer vision in Deep Learning.  

 

NOTE: This documentation applies to the MASTER version of DeepNet only.



## Install Dependencies



Install the required packages  

`pip install -r requirements.txt`



## Features



DeepNet currently supports the following features:



### Models



| Models | Description |

| ------- | ----- |

| [ResNet](./model/models/resnet.py) | ResNet-18 |

| [ResModNet](./model/models/resmodnet.py) | A modified version of ResNet-18  |

| [CustomNet](./model/models/customnet.py) | A modified version of ResNet-18   |

| [MaskNet3](./model/models/masknet.py) | A model to predict the Segmentation mask of the given image. |

| [DepthMaskNet8](./model/models/depthnet.py) | A model to predict the Monocular Depth Maps of the given image. |



### Training and Validation



| Functionality | Description |

| ------- | ----- |

| [Train](/home/rvk/DeepNet/model/train.py) | Training and Validation of the model |

| [Model](/home/rvk/DeepNet/model/learner.py) | Handles all the function for training a model  |

| [Dataset](/home/rvk/DeepNet/data/dataset) | Contains classes to handle data for training the model|



### Metrics



- [Mean Absolute Error](https://github.com/rvk007/DeepNet/blob/f67732d2d65798289925ea76d58f1d8636f13273/model/metrics.py#L36)

- [Root Mean Squared Error](https://github.com/rvk007/DeepNet/blob/f67732d2d65798289925ea76d58f1d8636f13273/model/metrics.py#L50)

- [Mean Absolute Relative Error](https://github.com/rvk007/DeepNet/blob/f67732d2d65798289925ea76d58f1d8636f13273/model/metrics.py#L67)

- [Intersection Over Union Error](https://github.com/rvk007/DeepNet/blob/f67732d2d65798289925ea76d58f1d8636f13273/model/metrics.py#L84)

- [Root Mean Square Error](https://github.com/rvk007/DeepNet/blob/f67732d2d65798289925ea76d58f1d8636f13273/model/metrics.py#L130)



### Losses



| Loss | Description |

| ------- | ----- |

| [Dice](./model/losses/dice_loss.py) | ResNet-18 |

| [SSIM](./model/losses/ssim.py) | A modified version of ResNet-18  |

| [MSE](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss.py#L5) | Mean squared error (squared L2 norm) between each element in the input and target   |

| [BCE](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss.py#L22) | Binary Cross Entropy between the target and the output |

| [BCEWithLogitsLoss](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss.py#L41) | Combination of Sigmoid layer and the BCE in one single class |

| [RMSE](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss.py#L58) | Root mean squared error (squared L2 norm) between each element in the input and target |



Weighted Combination of loss functions

- [BCE-RMSE](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss_combination.py#L10)

- [SSIM-RMSE](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss_combination.py#L22)

- [BCE-SSIM](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss_combination.py#L34)

- [RMSE-SSIM](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss_combination.py#L46)

- [SSIM-DICE](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss_combination.py#L57)

- [RMSE-DICE](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss_combination.py#L68)

- [BCE-DICE](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss_combination.py#L79)

- [RMSE-BCE-DICE](https://github.com/rvk007/DeepNet/blob/44ea35c02df7e719fc5c7f2f0c5da6f0cbfec4e3/model/losses/loss_combination.py#L103)



### Scheduler



- StepLR

- ReduceLROnPlateau

- OneCycleLR



### Data Augmentation



  - Resize

  - Padding

  - Random Crop

  - Horizontal Flip

  - Vertical Flip

  - Gaussian Blur

  - Random Rotation

  - CutOut



### Utilities



| Utility | Description |

| ------- | ----- |

| [GRADCAM](./gradcam/gradcam.py) | Calculates GradCAM(Gradient-weighted Class Activation Map) saliency map |

| [GradCAMpp](./gradcam/gradcam_pp.py) | Calculate GradCAM++ salinecy map using heatmap and image |

| [LRFinder](./lr_finder/lr_finder.py) | Range test to calculate optimal Learning Rate  |

| [Checkpoint](./utils/checkpoint.py) | Loading and saving checkpoints  |

| [ProgressBar](./utils/progress_bar.py) | Display Progress bar |

| [Tensorboard](./utils/tensorboard.py) | Creates Tensorboard visualization  |

| [Summary](./utils/summary.py)| Display model summary |

| [Plot](./utils/plot.py)| Plot the graph of a metric, prediction image and class accuracy |



## Dependencies



DeepNet has the following third-party dependencies



- numpy

- torch

- torchvision

- torchsummary

- tqdm

- matplotlib

- albumentations

- opencv-python



For a demo on how to use these modules, refer to the notebooks present in the [examples](./examples) directory.



## Contact/Getting Help



If you need any help or want to report a bug, raise an issue in the repo.