Practical overview
When it fits
Use Turi Create when a Python-based workflow for a listed machine-learning task and, where relevant, Core ML export matches your application needs and the documented environment requirements.
Who it is for
- Python developers adding machine-learning features to applications.
- Teams exploring labeled or other supported text, image, audio, video and sensor data.
- Developers preparing models for export to Core ML.
Documented uses
- Build recommendations that personalize choices for app users.
- Classify images, sounds, drawings, text or sensor activity.
- Detect objects in images or find similar images.
- Create models for regression, clustering, style transfer and other listed tasks.
- Generate predictions and export a model to Core ML.
Requirements & limits
The repository is public and lists a BSD-3-Clause license. The README recommends virtualenv and also documents an Anaconda environment route.
- The documented environment requires x86_64, at least 4 GB of RAM and Python 2.7 or Python 3.5–3.8.
- Windows support is specified as Windows 10 via WSL.
- A GPU is not required; acceleration is described only for selected models and supported GPU configurations.
- The repository is marked archived, an important consideration for adopting it as a new dependency.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install Turi Create in a Python virtual environment
Set up the package in an isolated Python environment using the documented virtualenv route.
Before you start
- A supported platform: macOS 10.12 or later, Linux with glibc 2.10 or later, or Windows 10 via WSL.
- x86_64 architecture, at least 4 GB of RAM, and Python 2.7 or Python 3.5–3.8.
Steps
Install virtualenv with pip.
Create an environment by running
virtualenv venv.Activate it with
source ~/venv/bin/activate.Install Turi Create with the command below.
pip install -U turicreate
Expected outcome
Turi Create is installed in the activated Python virtual environment.
Guide limitations
- The documented Windows route is Windows 10 via WSL.
- The repository is marked archived.
Build and export an image-classification model
Follow the documented example to create predictions from labeled image data and export the model to Core ML.
Before you start
- Turi Create installed in a supported Python environment.
- A labeled SFrame dataset in the documented format, with a
photoLabeltarget column.
Steps
Load the dataset with
tc.SFrame('photoLabel.sframe').Create an image classifier using
tc.image_classifier.create(data, target='photoLabel').Generate predictions with
model.predict(data).Export the model with
model.export_coreml('MyClassifier.mlmodel').
Expected outcome
The documented workflow produces predictions and exports a Core ML model file.
Guide limitations
- The example assumes the labeled dataset already exists.
- The README identifies Core ML as the export route; it does not document the steps for integrating the exported file into an application.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 12 October 2026
- Original website title
- apple/turicreate · Official homepage title · retained in its source language