A Guide for Model Production
This documentation is intended to aggregate resources from Roboflow's blog, YouTube channel, and other information from our team, to promote the success of all Roboflow users..
Key Metrics
mAP, Precision and Recall scores are provided for all models trained with Roboflow Train, and through the notebooks in the Roboflow Model Zoo.
mAP
- "When evaluating an object detection model in computer vision, mean average precision is the most commonly cited metric for assessing performance."
Precision
- A measure of how precise a model is at prediction time. True positives are divided by all positives that have been guessed.
Recall
- A measure of performance for a prediction system. Recall is used to assess whether a prediction system is guessing enough. True positives are divided by all possible true positives.
Prediction
- An attempt by a model to replicate the ground truth. A prediction usually contains a confidence value for each class.
Annotating Images for Object Detection
- Formats available for importing and exporting on Roboflow
Key Tools for Dataset Quality Control
Dataset Health Check
Dataset Health Check, Guided Walkthrough
Preprocessing and Augmentation
"Garbage in, garbage out."
This old machine learning adage conveys a salient machine learning point: unless input data is of high quality, model accuracy — even with the best computer vision architectures — will suffer. Understanding what preprocessing and augmentation are at their core enables data scientists to get the most out of their input data.
Preprocessing
Preprocessing steps are image changes that are done to all images, like resizing or converting to grayscale. Preprocessing steps will be applied to training, validation, and testing sets.
Augmentation
Image augmentation steps are image changes designed only to increase dataset size for better performance, like randomly altering brightness or rotation. Image augmentation steps will be applied only to training images.
Bounding Box Level Augmentation
Bounding box level augmentation generates new training data by only altering the content of a source image’s bounding boxes. In doing so, developers have greater control over creating training data that is more suitable to their problem’s conditions.
One-Click Model Training (Roboflow Train)
Third-Party Training Integrations (Pro)
AWS, Azure Custom Vision, Google Cloud Vision AutoML
- AWS & Azure Custom Vision require Third-Party keys to be entered into your project's workspace. These access points must be set up on your target AWS or Azure account to enable integration with Roboflow.
Zero Shot Object Tracking
Model Production Tips
Characteristics: Specific, Achievable, Measurable
- What will our model experience "in the wild…?" 🤔
Through the dataset assembly process, our laser focus might miss a whole host of edge cases that we would naturally consider out-of-scope for our model. While out-of-scope instances are intuitive to us, our model has no way of knowing anything beyond the scope of what it has been shown.
- Using pre-trained weights to bootstrap your model’s learning. You are “transferring” the knowledge learned on another dataset and then “fine-tuning” it to learn about your new domain.
Key Resources
Last modified 8mo ago
Copy link
Contents
Glossary of Common Computer Vision Terms
Key Metrics
mAP
Precision
Recall
Prediction
Annotating Images for Object Detection
Computer Vision Annotation Formats
Key Tools for Dataset Quality Control
Dataset Health Check
Preprocessing and Augmentation
"Garbage in, garbage out."
One-Click Model Training (Roboflow Train)
Introduction (Video)
Third-Party Training Integrations (Pro)
Zero Shot Object Tracking
Model Production Tips
Choosing the Right Problem Statement (Roboflow Blog)
The Ultimate Guide to Object Detection (Roboflow Blog)
Solving the Out of Scope Problem (Roboflow Blog) | Video
Augmentations for Aerial Datasets (Roboflow's Blog)
Small Object Detection (Roboflow's Blog) | Video
Active Learning Tips (Roboflow's Blog) | Video
A Primer on Transfer Learning (Roboflow's Blog)
Key Resources
Roboflow's YouTube Channel - All Videos
Roboflow's Blog
Roboflow's Featured Case Studies
Roboflow's Available Features