[NFC][mlir-tensorrt] Fix minor issues (#318)

This PR makes the following changes,
- Add code owners for `mir-tensorrt` part to help review.
- Add torch as dependency in `requirements-dev.txt`. Torch based test is
used in Python API.
- Fix an error in README.

.github/CODEOWNERS

@@ -1 +1,2 @@
 .github/workflows/* @parthchadha @pranavm-nvidia @jhalakpatel @yizhuoz004 @shelkesagar29
+/mlir-tensorrt/ @christopherbate @shelkesagar29

mlir-tensorrt/README.md

@@ -57,9 +57,9 @@ cmake --preset ninja-llvm
 ninja -C build all
 
 # Build and run tests
-ninja -C build/mlir-tensorrt check-mlir-executor
-ninja -C build/mlir-tensorrt check-mlir-tensorrt-dialect
-ninja -C build/mlir-tensorrt check-mlir-tensorrt
+ninja -C build check-mlir-executor
+ninja -C build check-mlir-tensorrt-dialect
+ninja -C build check-mlir-tensorrt
 
 # Build wheels (output in `build/wheels`)
 ninja -C build mlir-tensorrt-all-wheels

mlir-tensorrt/python/requirements-dev.txt

@@ -1,6 +1,7 @@
 # Torch is required by `optimum`. This will let PIP select the CPU version.
 # The GPU version installs `nvidia-` pip packages which confuse JAX.
 -r requirements.txt
+--extra-index-url https://download.pytorch.org/whl/cpu
 numpy<2.0.0
 click
 matplotlib
@@ -13,3 +14,4 @@ nvidia-ml-py
 black
 build
 psutil
+torch==2.1.0+cpu