Introduction¶
This toolchain captures TensorRT network creation and build parameters at runtime via a shim, then deterministically replays them to reproduce an engine build. Use it to debug or reproduce builds independent of the originating framework.
Prerequisites¶
TensorRT installed (ensure you know the absolute path to its
lib
andbin
directories)libtensorrt_shim.so
available in your TensorRTlib
directorytensorrt_player
available in your TensorRTbin
directory
Quick start: Capture¶
TORCHTRT_ENABLE_TENSORRT_API_CAPTURE=1 python test.py
You should see shim.json
and shim.bin
generated in /tmp/torch_tensorrt_{current_user}/shim
.
Replay: Build the engine from the capture¶
Use tensorrt_player
to replay the captured build without the original framework:
tensorrt_player -j /absolute/path/to/shim.json -o /absolute/path/to/output_engine
This produces a serialized TensorRT engine at output_engine
.
Validate the engine¶
Run the engine with trtexec
:
trtexec --loadEngine=/absolute/path/to/output_engine
Notes¶
Ensure the
libnvinfer.so
used by the shim matches the TensorRT version in your environment.If multiple TensorRT versions are installed, prefer absolute paths as shown above.
Currently, it is not supported to capture multiple engines, in case of graph break, only the first engine will be captured.