To run the demo Jupyter notebooks in this folder, follow the instructions in the TRT setup guide to build and launch the docker container, e.g. ./docker/build.sh --file docker/ubuntu-20.04.Dockerfile --tag tensorrt-ubuntu20.04-cuda11.7
and ./docker/launch.sh --tag tensorrt-ubuntu20.04-cuda11.7 --gpus all --jupyter <port>
by specifying the port number.
Then, use your browser to start the Jupyter lab interface by opening the token-protected link provided in the terminal, e.g. http://<host_name>:<port>/lab?token=...
.
Notebook list:
- gpt2.ipynb: Step by step walkthrough for building the GPT-2 TensorRT engine.
- gpt2-playground.ipynb: GUI for benchmarking GPT-2 TensorRT engines.
- t5.ipynb: Step by step walkthrough for building the T5 TensorRT engine.
- t5-playground.ipynb: GUI for benchmarking T5 TensorRT engines.
- bart.ipynb: Step by step walkthrough for building the BART TensorRT engine.
- bart-playground.ipynb: GUI for benchmarking BART TensorRT engines.