Skip to content

Latest commit

 

History

History
95 lines (88 loc) · 3.16 KB

CHANGELOG.md

File metadata and controls

95 lines (88 loc) · 3.16 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Version 0.1.0 - 2021-04-10

Modified

  • Updated License from MIT to GNU AGPL V3
  • Re-shuffling:
    • utils/data_loader.py -> core/data_loader.py
    • utils/model.py -> core/model.py
  • Moved Custom Callbacks to new file: utils/add_ons.py

Version 0.1.0 Beta - 2021-04-06

Added

  • Dockerfile
  • Launch Script (Dependency: GNU Parallel)
  • Automatic Tensorboard Initialization (Using the Launch Script) on Port 6006
  • Frontend:
    • Streamlit Dashboard For Easy Training and Visualization
    • Epoch Count and Batch Progress Bar with Training Status Message
    • Live Training & Validation Loss & Accuracy Plots on the dashboard using Plot.ly Graphs
    • Training and Validation Data Directory
    • Model Backbone Selector
    • Training Optimizer Selector
    • Learning Rate Slider
    • Batch Size Slider
    • Max Number of Epochs Selector
    • Input Image Shape Selector
    • Training Precision Selector
    • Training Button
    • Status Update with Final Validation Accuracy and Balloons Animation on Completion
  • Data Loader:
    • Optimized Tf.Data implementation for maximum GPU usage
    • Automatically handle errors such as corrupted images
    • Built-in Dataset Verification
    • Built-in Checks for if dataset is of a supported format
    • Supports Auto Detect Sub-folders get class information
    • Auto Generate Class Label Map
    • Built in Image Augmentation
    • Dataset Batch Visualization (With and Without Augment)
  • Model Trainer:
    • Support for Multiple Model Selection (All the models available to Keras)
    • Support for Loading Pre-Trained Model and Resume Training
    • Support for Mixed Precision Training for both GPUs and TPU optimized workloads
    • Support for Keras to Tensorflow SavedModel Converter
    • Contains a method to run Inference on a batch of input images
    • Dynamic Callbacks:
      • Automatic Learning Rate Decay based on validation accuracy
      • Automatic Training Stopping based on validation accuracy
      • Tensorboard Logging for Metrics
      • Autosave Best Model Weights at every epoch if validation accuracy increases
      • Support for any custom callbacks in addition to the above
    • Available Metrics (Training & Validation):
      • Categorical Accuracy
      • False Positives
      • False Negatives
      • Precision
      • Recall
      • Support for any custom metrics in addition to the above
  • Supported Models:
    • MobileNetV2
    • ResNet50V2
    • Xception
    • InceptionV3
    • VGG16
    • VGG19
    • ResNet50
    • ResNet101
    • ResNet152
    • ResNet101V2
    • ResNet152V2
    • InceptionResNetV2
    • DenseNet121
    • DenseNet169
    • DenseNet201
    • NASNetMobile
    • NASNetLarge
    • MobileNet
  • Supported Optimizers:
    • SGD
    • RMSprop
    • Adam
    • Adadelta
    • Adagrad
    • Adamax
    • Nadam
    • FTRL