fix: revert tensorrt version windows

Author: sangjanai

cpp/Makefile

@@ -24,8 +24,8 @@ endif
 # Build the Cortex engine
 build:
 ifeq ($(OS),Windows_NT)
-	@powershell -Command "cd ..; python .\scripts\build_wheel.py -a '80-real;86-real;89-real' --trt_root 'C:\workspace\TensorRT-10.1.0.27\' -D \"CMAKE_CUDA_COMPILER='C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.4/bin/nvcc.exe'\" -D 'BUILD_CORTEX_TENSORRT-LLM=ON' --use_ccache"
-	@powershell -Command "cd build; cmake .. -DCMAKE_CUDA_ARCHITECTURES='80-real;86-real;89-real' -DTRT_LIB_DIR='C:/workspace/TensorRT-10.1.0.27/lib' -DTRT_INCLUDE_DIR='C:/workspace/TensorRT-10.1.0.27/include' -DBUILD_CORTEX_TENSORRT-LLM=ON -DCMAKE_CUDA_COMPILER='C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.4/bin/nvcc.exe' -DENABLE_MULTI_DEVICE=0 -DCMAKE_CXX_COMPILER_LAUNCHER=ccache -DCMAKE_CUDA_COMPILER_LAUNCHER=ccache -G Ninja; cmake --build . --parallel 2 --config Release"
+	@powershell -Command "cd ..; python .\scripts\build_wheel.py -a '80-real;86-real;89-real' --trt_root 'C:\workspace\TensorRT-10.0.1.6\' -D \"CMAKE_CUDA_COMPILER='C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.4/bin/nvcc.exe'\" -D 'BUILD_CORTEX_TENSORRT-LLM=ON' --use_ccache"
+	@powershell -Command "cd build; cmake .. -DCMAKE_CUDA_ARCHITECTURES='80-real;86-real;89-real' -DTRT_LIB_DIR='C:/workspace/TensorRT-10.0.1.6/lib' -DTRT_INCLUDE_DIR='C:/workspace/TensorRT-10.0.1.6/include' -DBUILD_CORTEX_TENSORRT-LLM=ON -DCMAKE_CUDA_COMPILER='C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.4/bin/nvcc.exe' -DENABLE_MULTI_DEVICE=0 -DCMAKE_CXX_COMPILER_LAUNCHER=ccache -DCMAKE_CUDA_COMPILER_LAUNCHER=ccache -G Ninja; cmake --build . --parallel 2 --config Release"
 else
 	@cd .. && python3 ./scripts/build_wheel.py --trt_root /usr/local/tensorrt --cuda_architectures "80-real;86-real;89-real" --extra-cmake-vars "BUILD_CORTEX_TENSORRT-LLM=ON" --use_ccache
 	@cd build && cmake .. -DCMAKE_CUDA_ARCHITECTURES='80-real;86-real;89-real' -DTRT_LIB_DIR='/usr/local/tensorrt/lib' -DTRT_INCLUDE_DIR='/usr/local/tensorrt/include' -DBUILD_CORTEX_TENSORRT-LLM=ON -DCMAKE_CXX_COMPILER_LAUNCHER=ccache -DCMAKE_CUDA_COMPILER_LAUNCHER=ccache -G Ninja && cmake --build . --config Release;
@@ -35,7 +35,7 @@ endif
 pre-package:
 ifeq ($(OS),Windows_NT)
 	@powershell -Command "cd tensorrt_llm\cortex.tensorrt-llm\; mkdir -p cortex.tensorrt-llm; cp ..\..\build\tensorrt_llm\cortex.tensorrt-llm\engine.dll cortex.tensorrt-llm\;"
-	@powershell -Command "cd tensorrt_llm\cortex.tensorrt-llm\; cp -Force C:\workspace\TensorRT-10.1.0.27\lib\nvinfer_10.dll cortex.tensorrt-llm\;"
+	@powershell -Command "cd tensorrt_llm\cortex.tensorrt-llm\; cp -Force C:\workspace\TensorRT-10.0.1.6\lib\nvinfer_10.dll cortex.tensorrt-llm\;"
 	@powershell -Command "cd tensorrt_llm\cortex.tensorrt-llm\; cp -Force ..\..\build\tensorrt_llm\plugins\nvinfer_plugin_tensorrt_llm.dll cortex.tensorrt-llm\;"
 	@powershell -Command "cd tensorrt_llm\cortex.tensorrt-llm\; cp -Force ..\..\build\tensorrt_llm\tensorrt_llm.dll cortex.tensorrt-llm\;"
 	@powershell -Command "cd tensorrt_llm\cortex.tensorrt-llm\; cp -Force .\build_deps\_install\bin\zlib.dll cortex.tensorrt-llm\;"

docs/source/installation/build-from-source-windows.md

@@ -11,7 +11,7 @@ This section is for advanced users. Skip this section if you plan to use the pre
 1. Install prerequisites listed in our [Installing on Windows](https://nvidia.github.io/TensorRT-LLM/installation/windows.html) document.
 2. Install [CMake](https://cmake.org/download/), version 3.27.7 is recommended, and select the option to add it to the system path.
 3. Download and install [Visual Studio 2022](https://visualstudio.microsoft.com/).
-4. Download and unzip [TensorRT 10.1.0.27](https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.1.0/zip/TensorRT-10.1.0.27.Windows10.win10.cuda-12.4.zip).
+4. Download and unzip [TensorRT 10.0.1.6](https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.0.1/zip/TensorRT-10.0.1.6.Windows10.win10.cuda-12.4.zip).
 
 ## Building a TensorRT-LLM Docker Image
 
@@ -65,7 +65,7 @@ git submodule update --init --recursive
 2. Build TensorRT-LLM. This command generates `build\tensorrt_llm-*.whl`.
 
 ```bash
-python .\scripts\build_wheel.py -a "89-real" --trt_root C:\workspace\TensorRT-10.1.0.27\
+python .\scripts\build_wheel.py -a "89-real" --trt_root C:\workspace\TensorRT-10.0.1.6\
 ```
 
 3. Copy or move `build\tensorrt_llm-*.whl` into your mounted folder so it can be accessed on your host machine. If you intend to use the C++ runtime, you'll also need to gather various DLLs from the build into your mounted folder. For more information, refer to [C++ Runtime Usage](#c-runtime-usage).
@@ -103,7 +103,7 @@ python .\scripts\build_wheel.py -a "89-real" --trt_root C:\workspace\TensorRT-10
 
         1. Install [CMake](https://cmake.org/download/), version 3.27.7 is recommended, and select the option to add it to the system path.
         2. Download and install [Visual Studio 2022](https://visualstudio.microsoft.com/). When prompted to select more Workloads, check **Desktop development with C++**.
-        3. Download and unzip [TensorRT 10.1.0.27](https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.1.0/zip/TensorRT-10.1.0.27.Windows10.win10.cuda-12.4.zip). Move the folder to a location you can reference later, such as `%USERPROFILE%\inference\TensorRT`.
+        3. Download and unzip [TensorRT 10.0.1.6](https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.0.1/zip/TensorRT-10.0.1.6.Windows10.win10.cuda-12.4.zip). Move the folder to a location you can reference later, such as `%USERPROFILE%\inference\TensorRT`.
 
             1. Add the libraries for TensorRT  to your system's `Path` environment variable. Your `Path` should include a line like this:
 

docs/source/release-notes.md

@@ -8,7 +8,7 @@ All published functionality in the Release Notes has been fully tested and verif
 ## TensorRT-LLM Release 0.10.0
 
 ### Announcements
-- TensorRT-LLM supports TensorRT 10.1.0 and NVIDIA NGC 24.03 containers.
+- TensorRT-LLM supports TensorRT 10.0.1 and NVIDIA NGC 24.03 containers.
 
 ### Key Features and Enhancements
 - The Python high level API
@@ -81,7 +81,7 @@ All published functionality in the Release Notes has been fully tested and verif
 ### Infrastructure changes
   - Base Docker image for TensorRT-LLM is updated to `nvcr.io/nvidia/pytorch:24.03-py3`.
   - Base Docker image for TensorRT-LLM backend is updated to `nvcr.io/nvidia/tritonserver:24.03-py3`.
-  - The dependent TensorRT version is updated to 10.1.0.
+  - The dependent TensorRT version is updated to 10.0.1.
   - The dependent CUDA version is updated to 12.4.0.
   - The dependent PyTorch version is updated to 2.2.2.
 

requirements-windows.txt

@@ -15,7 +15,7 @@ h5py==3.10.0
 pywin32
 StrEnum
 sentencepiece>=0.1.99
-tensorrt==10.1.0
+tensorrt==10.0.1
 tokenizers>=0.14
 # Default torch is CPU-only on Windows, so need to specify a torch version with GPU support
 torch @ https://download.pytorch.org/whl/cu121/torch-2.2.0%2Bcu121-cp310-cp310-win_amd64.whl#sha256=8f54c647ee19c8b4c0aad158c73b83b2c06cb62351e9cfa981540ce7295a9015

windows/setup_build_env.ps1

@@ -45,21 +45,21 @@ if (-not $skipVSBuildTools) {
     Write-Output "Skipping Visual Studio Build Tools installation"
 }
 
-# Install TensorRT 10.1.0 for TensorRT-LLM
+# Install TensorRT 10.0.1 for TensorRT-LLM
 if (-not $skipTRT) {
     Write-Output "Downloading TensorRT"
-    Invoke-WebRequest -Uri 'https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.1.0/zip/TensorRT-10.1.0.27.Windows10.win10.cuda-12.4.zip' -OutFile 'TensorRT-10.1.0.27.zip'
+    Invoke-WebRequest -Uri 'https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.0.1/zip/TensorRT-10.0.1.6.Windows10.win10.cuda-12.4.zip' -OutFile 'TensorRT-10.0.1.6.zip'
     Write-Output "Extracting TensorRT"
     # Get path
     $absolutePath = Resolve-Path $TRTPath
-    Expand-Archive -Path '.\TensorRT-10.1.0.27.zip' -DestinationPath $absolutePath
+    Expand-Archive -Path '.\TensorRT-10.0.1.6.zip' -DestinationPath $absolutePath
     Write-Output "Removing TensorRT zip"
-    Remove-Item -Path 'TensorRT-10.1.0.27.zip' -Force
+    Remove-Item -Path 'TensorRT-10.0.1.6.zip' -Force
     Write-Output "Adding TensorRT to system Path"
-    [Environment]::SetEnvironmentVariable('Path', "$env:Path;$absolutePath\TensorRT-10.1.0.27\lib", [EnvironmentVariableTarget]::Machine)
+    [Environment]::SetEnvironmentVariable('Path', "$env:Path;$absolutePath\TensorRT-10.0.1.6\lib", [EnvironmentVariableTarget]::Machine)
     Write-Output "Installing TensorRT Python wheel"
-    python3 -m pip install $absolutePath\TensorRT-10.1.0.27\python\tensorrt-10.1.0-cp310-none-win_amd64.whl
-    Write-Output "Done TensorRT installation at '$absolutePath\TensorRT-10.1.0.27'"
+    python3 -m pip install $absolutePath\TensorRT-10.0.1.6\python\tensorrt-10.0.1-cp310-none-win_amd64.whl
+    Write-Output "Done TensorRT installation at '$absolutePath\TensorRT-10.0.1.6'"
 } else {
     Write-Output "Skipping TensorRT installation"
 }

windows/setup_env.ps1

@@ -176,10 +176,10 @@ if (-not ($skipTRT)) {
         Write-Output "Grabbing TensorRT..."
         $ProgressPreference = 'SilentlyContinue'
         New-Item -Path .\TensorRT -ItemType Directory
-        Invoke-WebRequest -Uri 'https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.1.0/zip/TensorRT-10.1.0.27.Windows10.win10.cuda-12.4.zip' -OutFile .\TensorRT\trt.zip
+        Invoke-WebRequest -Uri 'https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.0.1/zip/TensorRT-10.0.1.6.Windows10.win10.cuda-12.4.zip' -OutFile .\TensorRT\trt.zip
         Expand-Archive -Path .\TensorRT\trt.zip -DestinationPath .\TensorRT\
         Remove-Item -Path .\TensorRT\trt.zip -Force
-        $trtPath = Join-Path $TRT_BASE TensorRT-10.1.0.27
+        $trtPath = Join-Path $TRT_BASE TensorRT-10.0.1.6
         Write-Output "TensorRT installed at ${trtPath}"
 
         $trtSubPaths = @{