# qqwweee/keras-yolo3

**Attribution required: if you use, quote, or summarise this content, you must credit and link back to [awesome-repositories.com](https://awesome-repositories.com/repository/qqwweee-keras-yolo3).**

7,116 stars · 3,381 forks · Python · MIT

## Links

- GitHub: https://github.com/qqwweee/keras-yolo3
- awesome-repositories: https://awesome-repositories.com/repository/qqwweee-keras-yolo3.md

## Description

This project is an object detection framework implementing the YOLOv3 architecture using Keras and TensorFlow. It functions as a deep learning vision model and computer vision toolset designed to locate and classify multiple entities within images and video streams using bounding boxes.

The system includes a multi-GPU inference engine to distribute computational loads across several graphics processing units. It also provides a pipeline for creating custom object detectors by retraining pre-trained weights on annotated datasets to recognize user-defined object classes.

The framework covers model training and fine-tuning through a two-stage retraining process and weight optimization. It includes utilities for network architecture configuration via external files, weight format conversion between frameworks, and the transformation of VOC annotations into plain text for training.

The project supports inference across static images, live streams, and sequential video files.

## Tags

### Part of an Awesome List

- [YOLOv3 Implementations](https://awesome-repositories.com/f/awesome-lists/ai/object-detection-frameworks/yolov3-implementations.md) — Implements the YOLOv3 architecture using Keras and TensorFlow for real-time image and video object detection.
- [Deep Learning Models](https://awesome-repositories.com/f/awesome-lists/ai/deep-learning-models.md) — Implements a deep learning vision model for locating and classifying multiple entities using bounding boxes.
- [Vision Model Fine-Tuning](https://awesome-repositories.com/f/awesome-lists/ai/model-training-and-fine-tuning/vision-model-fine-tuning.md) — Optimizes the final layers of pre-trained vision models to recognize specific custom object classes. ([source](https://github.com/qqwweee/keras-yolo3/blob/master/train_bottleneck.py))
- [Computer Vision Libraries](https://awesome-repositories.com/f/awesome-lists/ai/computer-vision-libraries.md) — Keras implementation of the YOLOv3 object detector.

### Artificial Intelligence & ML

- [Object Detection](https://awesome-repositories.com/f/artificial-intelligence-ml/computer-vision-systems/computer-vision/object-detection-tracking/object-detection.md) — Identifies and locates multiple objects within static images using bounding boxes and classification. ([source](https://github.com/qqwweee/keras-yolo3/blob/master/yolo_video.py))
- [Video Stream Detections](https://awesome-repositories.com/f/artificial-intelligence-ml/computer-vision-systems/computer-vision/object-detection-tracking/object-detection/video-stream-detections.md) — Processes sequential video frames to detect objects and exports the resulting detection output. ([source](https://github.com/qqwweee/keras-yolo3/blob/master/yolo_video.py))
- [Real-Time Object Detection](https://awesome-repositories.com/f/artificial-intelligence-ml/computer-vision-systems/computer-vision/object-detection-tracking/real-time-object-detection.md) — Provides real-time identification and localization of objects within live images and video streams. ([source](https://github.com/qqwweee/keras-yolo3/blob/master/yolov3-tiny.cfg))
- [Detection Model Training](https://awesome-repositories.com/f/artificial-intelligence-ml/custom-model-training/detection-model-training.md) — Allows training object detection models on custom annotated datasets using pre-trained weights. ([source](https://github.com/qqwweee/keras-yolo3#readme))
- [Deep Learning Inference Engines](https://awesome-repositories.com/f/artificial-intelligence-ml/deep-learning-inference-engines.md) — Executes pretrained neural networks on visual data to generate bounding boxes and class labels.
- [Keras Model Implementations](https://awesome-repositories.com/f/artificial-intelligence-ml/neural-network-implementations/keras-model-implementations.md) — Implements the YOLOv3 architecture specifically using the high-level Keras API.
- [Computer Vision Training Frameworks](https://awesome-repositories.com/f/artificial-intelligence-ml/computer-vision-systems/computer-vision/development-orchestration-tools/computer-vision-training-frameworks.md) — Provides a toolset for training, fine-tuning, and deploying convolutional neural networks for visual recognition.
- [Vision Model Retraining Pipelines](https://awesome-repositories.com/f/artificial-intelligence-ml/instruction-fine-tuning/multi-stage-fine-tuning-frameworks/two-stage-fine-tuning-pipelines/vision-model-retraining-pipelines.md) — Employs a two-stage retraining process that first stabilizes frozen layers before training the entire network. ([source](https://github.com/qqwweee/keras-yolo3/blob/master/train.py))
- [Multi-GPU Inference Runtimes](https://awesome-repositories.com/f/artificial-intelligence-ml/model-inference-runtimes/multi-gpu-inference-runtimes.md) — Distributes the computational load of deep learning predictions across multiple graphics processing units.
- [Multi-GPU Workload Distribution](https://awesome-repositories.com/f/artificial-intelligence-ml/model-optimization/inference-deployment/model-deployment-toolkits/distributed-deployment-utilities/multi-gpu-workload-distribution.md) — Distributes the computational load across multiple GPUs to accelerate object detection processing speed.
- [Two-Stage Weight Stabilization](https://awesome-repositories.com/f/artificial-intelligence-ml/supervised-model-weight-optimization/two-stage-weight-stabilization.md) — Optimizes the model through a two-stage retraining process that stabilizes early layers before full-network fine-tuning.
- [Video Analysis](https://awesome-repositories.com/f/artificial-intelligence-ml/video-analysis.md) — Processes video files to automatically detect and track entities across frames for monitoring or tagging.
- [Custom YOLO Detectors](https://awesome-repositories.com/f/artificial-intelligence-ml/video-object-tracking/yolo-object-detectors/custom-yolo-detectors.md) — Provides a pipeline for retraining pre-trained YOLO weights on custom datasets to recognize user-defined object classes.

### Software Engineering & Architecture

- [Model Architecture Configurations](https://awesome-repositories.com/f/software-engineering-architecture/application-lifecycle-management/configuration-management/automation-and-templating-frameworks/configuration-modularization/model-architecture-configurations.md) — Allows defining the neural network structure using external configuration files instead of hard-coded layers.