convertor_tf_onnx_trt

tensorflow -> onnx -> tensorRT

Build the image

• Follow https://stackoverflow.com/a/61737404 first
• When it is building, it will automatically run a model conversion tensorflow -> onnx -> tensorrt, then infer an image. Check the dockerfile for more details.
• Run

  docker build -t neilvaltec/convertor_tf_onnx_trt:0.0.1 .

Run the contianer

docker run -it --rm --gpus all neilvaltec/convertor_tf_onnx_trt:0.0.1 bash

Environments

• Cloud provider
  • AWS
• Image
  • amazon/Deep Learning AMI GPU CUDA 11.2.1 (Ubuntu 20.04) 20220607
• Instance type
  • g4dn.2xlarge (NVIDIA T4 GPU * 1)