SP

Sahil P.

AIML Engineer

Mid-Market (51-1000 emp.)

10/7/2025

"Edge devices support is Incredible"

5/5

What do you like best about Ultralytics?

We can easily train on our custom dataset and then export an optimized Onnx, TensorRT(engine) and Int8 model. This makes deployment on our edge devices like the Jetson ORIN incredibly efficient. Review collected by and hosted on G2.com.

What do you dislike about Ultralytics?

While generally good, the documentation for very specific, advanced deployment scenarios can be lacking.

I have had issues with RTSP pipeline with h264/265 codec supports. Review collected by and hosted on G2.com.

What problems is Ultralytics solving and how is that benefiting you?

It provides a high-level API and command-line interface that abstracts away this complexity, I can immediately focus on the most important part of your project: curating a robust dataset.

To get real-time performance on edge hardware like a Jetson ORIN, the model must be converted into a highly optimized format like TensorRT, which is a difficult and error-prone process, Ultralytics has this resolved, which is a massive benefit for my specific use case. Review collected by and hosted on G2.com.

Show More

Current UserValidated ReviewerIncentivizedSource: Seller invite

See what 2 reviewers think of Ultralytics

5.0 out of 5 · Verified reviews from real users

[
Read all reviews
](https://www.g2.com/products/ultralytics/reviews)