Students, Hobbyists or Individuals Training a Cat Detection Model
Machine Learning Engineers, Researchers, and Entrepreneurs
Startups, Research Laboratories, and Advanced Use Cases
We understand that companies have varying requirements and usage volume. If you have a different idea on how to use Datature, contact us.
No CUDA installation whatsoever, use only your favourite browser.
Annotate with your team or import existing annotation in real-time.
From bounding-box to masks, train on industry-standard models.
Manage and perform transfer learning on existing checkpoints.
Perform image augmentations and data transforms WITHOUT CODE.
Setup complex pipelines in minutes and no code = no bugs.
When you create a workflow and set the corresponding parameters, a private compute instance will be created with the transpiled TensorFlow codes. This instance will perform the training on your image data and you will be billed on a per minute basis.
We have a bunch of demos and guides over at docs.datature.io! We will be launching a how-to video series explaining our product. That said, you can also reach out to us via our community slack to speak with us as well as our developers. Finally, if you wish to discuss enterprise use cases or projects, just email us at firstname.lastname@example.org!
We support JPG and PNG image data format on our platform currently. However, if you have specialized data formats like DICOM or TIFF, you can convert them to either JPGs or PNGs prior to uploading.
We currently do not support video files, but it is a feature we are working hard on!
All labeled data, image assets, and user information is encrypted on Datature's server. We currently use Google Cloud Platform for cloud storage where data is encrypted with GCP's encryption keys and a unique signed url is generated whenever a request is made. Additionally, we do not store any payment information such as credit cards or addresses on our server.
If you'd like to upgrade your plan or cancel an existing plan - please email email@example.com