DEPLOY NOW.
FLEXIBLE.EFFICIENT.EASY.

Test our deployment software for any model from common frameworks and on different target hardware.
Test now

How it works

1
Choose a sample model or upload your own model.
2
We compile the model and run it on the selected target hardware.
3
We share the key results with you: latency and peak memory in comparison to industry standard (TensorFlow Lite and TorchInductor).

please Choose

sample model
own model
next

model input

MobileNetV4-Conv-Large

Hardware input

Check your target hardware
back
 next

You're all set!

Where should we send the results?
Please see our Privacy Policy.
back
Thanks for submitting the form. Let us do the magic and we will come back to you as soon as possible.
back to the demo page
Oops! Something went wrong while submitting the form.

Choose model

back
 next

Model Requirements

  • File size: <100 MB
  • File type: .tflite, .pt, .pth, .onnx
Roofline declares to not reuse, reverse engineer, or steal any of your information or your model, and will not sell or provide any data to third parties without your approval.

model input

arg_0
Datatype
arg_1
Datatype
arg_3
Datatype
arg_4
Datatype
arg_5
Datatype
Click to add another argument

Hardware input

Check your target hardware
back
 next

You're all set!

Where should we send the results?
Please see our Privacy Policy.
back
Thanks for submitting the form. Let us do the magic and we will come back to you as soon as possible.
back to the demo page
Oops! Something went wrong while submitting the form.

See next-gen deployment in action

We understand the challenges of bringing fast-evolving applications to diverse and complex hardware.
Whether you have questions on how it works, want our opinion on a specific model you are aiming to deploy, or would like to explore our next-gen compiler for your product, we are here to help.
Reach out
Imprint Privacy Policy
Contact Us Email LinkedIn