Roboflow + NVIDIA TRT
Roboflow provides a Docker container for use with TensorRT. This container supports inference using Roboflow models for object detection, classification, and instance segmentation tasks.
Roboflow enterprise customers can use the Roboflow TensorRT Docker container for inference. To use the inference server, you will need to download the nvidia-container-runtime environment. View the full TensorRT documentation for more information. You can run inference for the following tasks:
- Object detection;
- Instance segmentation;
- Image classification and;
- OpenAI CLIP.
