Overview of Turi Create
Turi Create provides a simplified approach to developing custom machine learning models. It is especially useful for those who may not have an extensive background in machine learning, allowing them to integrate functionalities such as recommendations, object detection, image classification, image similarity, or activity classification into their applications seamlessly.
Key Features
- Easy-to-use: Turi Create emphasizes task-oriented development rather than focusing on specific algorithms, making the process more accessible to users.
- Visual: It offers built-in, streaming visualizations that help users explore and understand their data more effectively.
- Flexible: Capable of handling a variety of data types including text, images, audio, video, and sensor data.
- Fast and Scalable: Enables efficient work with large datasets on a single machine, providing scalability.
- Ready To Deploy: The models created can be exported to Core ML, which is compatible with iOS, macOS, watchOS, and tvOS apps.
Machine Learning Tasks Achievable with Turi Create
Turi Create supports a wide range of machine learning tasks, enabling users to perform:
- Recommender Systems: Enhances personalization by tailoring user choices.
- Image Classification: Assigns labels to images.
- Drawing Classification: Identifies drawings and gestures made with pencils or touch.
- Sound Classification: Categorizes different sounds.
- Object Detection: Detects and recognizes objects within images.
- One Shot Object Detection: Identifies 2D objects using a single example in images.
- Style Transfer: Applies artistic styles to images.
- Activity Classification: Detects activities using sensor data.
- Image Similarity: Finds and relates similar images.
- Classifiers: Predicts labels for data points.
- Regression: Forecasts numeric values.
- Clustering: Groups similar data points.
- Text Classification: Analyzes and interprets message sentiment.
Example: Building an Image Classifier
Creating an image classifier with Turi Create requires just a few lines of code:
import turicreate as tc
# Load data
data = tc.SFrame('photoLabel.sframe')
# Create a model
model = tc.image_classifier.create(data, target='photoLabel')
# Make predictions
predictions = model.predict(data)
# Export to Core ML
model.export_coreml('MyClassifier.mlmodel')
This simplified process allows developers to implement image recognition capabilities into iOS applications with ease.
Supported Platforms
Turi Create is versatile in platform support, compatible with:
- macOS 10.12 and above
- Linux (requires glibc 2.10 or later)
- Windows 10 (using Windows Subsystem for Linux)
System Requirements
To effectively run Turi Create, the following system requirements must be met:
- Python versions: 2.7, 3.5, 3.6, 3.7, 3.8
- x86_64 architecture
- A minimum of 4 GB of RAM
Installation
Turi Create can be installed following standard Python package installation procedures. Utilizing virtual environments such as virtualenv or Anaconda's conda can streamline and manage installations more efficiently.
For creating a Python virtual environment and installing Turi Create, use:
# Create a Python virtual environment
cd ~
virtualenv venv
# Activate your virtual environment
source ~/venv/bin/activate
# Install Turi Create
(venv) pip install -U turicreate
Documentation and Support
Comprehensive guides and API documentation are available to assist users in making the most out of Turi Create. Additionally, GPU support is available, significantly enhancing processing speeds for certain models, especially on macOS and Linux systems.
Community and Contributions
Turi Create encourages an open and inclusive community. Contributions are welcome, adhering to the project's code of conduct and contribution guidelines to ensure a positive collaborative environment.