Add Intel XPU support

README.md

@@ -64,6 +64,9 @@ CUDA **12.8** is recommended.
 git clone https://github.com/tensorforger/FluxRT
 cd FluxRT
 "scripts/install.bat"
+
+# PyTorch with Intel (XPU) support — use this instead of the line above
+# "scripts/install_xpu.bat"
 ```
 
 GUI reqires [OBS](https://obsproject.com/download) to be installed to access virtual webcam. 
@@ -74,6 +77,9 @@ GUI reqires [OBS](https://obsproject.com/download) to be installed to access vir
 git clone https://github.com/tensorforger/FluxRT
 cd FluxRT
 sh scripts/install.sh
+
+# PyTorch with Intel (XPU) support — use this instead of the line above
+# ./scripts/install_xpu.sh
 ```
 
 GUI reqires **v4l2loopback** to be installed and loaded to access virtual webcam. 
@@ -96,9 +102,12 @@ cd FluxRT
 conda create -n fluxrt python=3.12 pip -y
 conda activate fluxrt
 
-# Install PyTorch with CUDA support (adjust if needed)
+# PyTorch with CUDA support
 pip install torch torchvision --index-url https://download.pytorch.org/whl/cu128
 
+# PyTorch with Intel (XPU) support — use this instead of the line above
+# pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/xpu
+
 # Install project dependencies
 pip install -r requirements.txt
 pip install -e .
@@ -110,10 +119,15 @@ pip install -e .
 # Install uv if you don't have it
 pip install uv
 
-# Create environment and install PyTorch with CUDA support
+# Create environment
 uv venv --python 3.12
+
+# PyTorch with CUDA support
 uv pip install torch torchvision --index-url https://download.pytorch.org/whl/cu128
 
+# PyTorch with Intel (XPU) support — use this instead of the line above
+# uv pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/xpu
+
 # Install project dependencies
 uv pip install -r requirements.txt
 uv pip install -e .
@@ -225,7 +239,14 @@ FluxRT/
 
 ## 4. Run
 
-To enable `int8` quantization you can either add falg `--int8` when running any script or set `enable_int8_quantization` to `true` in  the corresponding config.
+To enable `int8` quantization you can either add flag `--int8` when running any script or set `enable_int8_quantization` to `true` in the corresponding config.
+
+To select the compute device, set `"device"` in the config JSON:
+
+```json
+"device": "cuda"   // NVIDIA GPU (default)
+"device": "xpu"    // Intel GPU
+```
 
 Run any script with conda environment activated:
 

configs/benchmark_config.json

@@ -9,6 +9,7 @@
         "width" : 576
     },
     "compile_models": true,
+    "device": "cuda",
     "enable_spatial_cache": true,
     "enable_int8_quantization": false,
     "target_fps":null,

configs/config_with_reference.json

@@ -9,6 +9,7 @@
         "width" : 576
     },
     "compile_models": true,
+    "device": "cuda",
     "enable_spatial_cache": true,
     "enable_int8_quantization": false,
     "target_fps":null,

configs/paint_config.json

@@ -9,6 +9,7 @@
         "width" : 576
     },
     "compile_models": false,
+    "device": "cuda",
     "enable_spatial_cache": true,
     "enable_int8_quantization": false,
     "target_fps":null,

configs/stream_processor_config.json

@@ -9,6 +9,7 @@
         "width" : 576
     },
     "compile_models": true,
+    "device": "cuda",
     "enable_spatial_cache": true,
     "enable_int8_quantization": false,
     "target_fps":null,

scripts/install_xpu.bat

@@ -0,0 +1,211 @@
+@echo off
+SETLOCAL EnableDelayedExpansion
+
+:: FluxRT installation script for Windows.
+:: Run from the repository root: scripts\install.bat
+
+:: ── sanity-check: running from repo root ──────────────────────────────────────
+IF NOT EXIST "pyproject.toml" (
+    echo [ERROR] This script must be run from the FluxRT repository root.
+    exit /b 1
+)
+
+:: ── prerequisites ─────────────────────────────────────────────────────────────
+echo [+] Checking prerequisites...
+
+where git >nul 2>&1
+IF ERRORLEVEL 1 (
+    echo [ERROR] 'git' is not installed. Install it from https://git-scm.com/download/win
+    exit /b 1
+)
+
+SET UV_DIR=thirdparty
+SET UV=%UV_DIR%\uv.exe
+
+IF EXIST "%UV%" (
+    echo [+] uv already present at '%UV%'.
+) ELSE (
+    echo [!] uv not found. Downloading to '%UV_DIR%'...
+    IF NOT EXIST "%UV_DIR%" mkdir "%UV_DIR%"
+    powershell -NoProfile -ExecutionPolicy Bypass -Command ^
+        "$zip = [System.IO.Path]::GetTempFileName() + '.zip';" ^
+        "Invoke-WebRequest -Uri 'https://github.com/astral-sh/uv/releases/latest/download/uv-x86_64-pc-windows-msvc.zip' -OutFile $zip;" ^
+        "Expand-Archive $zip -DestinationPath '%UV_DIR%' -Force;" ^
+        "Remove-Item $zip"
+    IF ERRORLEVEL 1 (
+        echo [ERROR] Failed to download uv. Check your internet connection and re-run.
+        exit /b 1
+    )
+    IF NOT EXIST "%UV%" (
+        echo [ERROR] uv.exe not found after extraction. Re-run the script.
+        exit /b 1
+    )
+    echo [+] uv installed successfully to '%UV_DIR%'.
+)
+
+git lfs version >nul 2>&1
+IF ERRORLEVEL 1 (
+    echo [ERROR] 'git-lfs' is not installed.
+    echo         Install with: winget install GitHub.GitLFS
+    echo         Or download from https://git-lfs.com
+    exit /b 1
+)
+
+echo [+] All prerequisites found.
+
+:: ── uv virtual environment ───────────────────────────────────────────────────
+SET VENV_DIR=fluxrt
+
+IF EXIST "%VENV_DIR%" (
+    echo [+] Virtual environment '%VENV_DIR%' already exists.
+) ELSE (
+    echo [+] Creating virtual environment '%VENV_DIR%' with python=3.12 via uv...
+    "%UV%" venv "%VENV_DIR%" --python 3.12
+    IF ERRORLEVEL 1 (
+        echo [ERROR] Failed to create virtual environment via uv.
+        exit /b 1
+    )
+)
+
+CALL "%VENV_DIR%\Scripts\activate.bat"
+IF ERRORLEVEL 1 (
+    echo [ERROR] Failed to activate virtual environment '%VENV_DIR%'.
+    exit /b 1
+)
+
+:: ── PyTorch ───────────────────────────────────────────────────────────────────
+python -c "import torch" >nul 2>&1
+IF ERRORLEVEL 1 (
+    echo [+] Installing PyTorch with XPU support...
+    "%UV%" pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/xpu
+    IF ERRORLEVEL 1 (
+        echo [ERROR] Failed to install PyTorch.
+        exit /b 1
+    )
+) ELSE (
+    echo [+] PyTorch is already installed.
+)
+
+:: ── Python requirements ───────────────────────────────────────────────────────
+:: Use 'diffusers' as a proxy — it's the heaviest transitive dependency.
+python -c "import diffusers" >nul 2>&1
+IF ERRORLEVEL 1 (
+    echo [+] Installing Python requirements from requirements.txt...
+    "%UV%" pip install -r requirements.txt
+    IF ERRORLEVEL 1 (
+        echo [ERROR] Failed to install requirements.
+        exit /b 1
+    )
+) ELSE (
+    echo [+] Python requirements already installed.
+)
+
+:: ── triton (Intel XPU backend) ───────────────────────────────────────────────
+:: Intel® XPU Backend for Triton* support for Windows is currently experimental and requires compiling PyTorch and Intel® XPU Backend for Triton* from source.
+:: Follow the Windows build guide:
+:: https://github.com/intel/intel-xpu-backend-for-triton/blob/main/.github/WINDOWS.md
+echo.
+echo ============================================================
+echo WARNING: Triton is required for model compilation
+echo ============================================================
+echo.
+echo     Intel XPU Backend for Triton support on Windows is
+echo     currently experimental and requires compiling PyTorch
+echo     and Intel XPU Backend for Triton from source.
+echo.
+echo     Windows build guide:
+echo     https://github.com/intel/intel-xpu-backend-for-triton/blob/main/.github/WINDOWS.md
+echo.
+echo ============================================================
+echo.
+
+:: ── fluxrt package ────────────────────────────────────────────────────────────
+python -c "from fluxrt import StreamProcessor" >nul 2>&1
+IF ERRORLEVEL 1 (
+    echo [+] Installing fluxrt package in editable mode...
+    "%UV%" pip install -e .
+    IF ERRORLEVEL 1 (
+        echo [ERROR] Failed to install fluxrt package.
+        exit /b 1
+    )
+) ELSE (
+    echo [+] fluxrt package already installed.
+)
+
+:: ── model downloads ───────────────────────────────────────────────────────────
+:: Register LFS hooks for the current user (idempotent).
+git lfs install
+
+:: ── RIFE frame-interpolation model ───────────────────────────────────────────
+SET RIFE_DIR=RIFE-safetensors
+SET RIFE_SENTINEL=RIFE-safetensors\flownet.safetensors
+IF EXIST "%RIFE_SENTINEL%" (
+    echo [+] RIFE frame-interpolation model: already downloaded.
+) ELSE IF EXIST "%RIFE_DIR%\.git" (
+    echo [!] RIFE: directory exists but looks incomplete — resuming LFS download...
+    git -C "%RIFE_DIR%" pull --ff-only
+    git -C "%RIFE_DIR%" lfs pull
+) ELSE IF EXIST "%RIFE_DIR%" (
+    echo [!] RIFE: '%RIFE_DIR%' exists but is not a git repository.
+    echo [!]       Remove it and re-run to download the model.
+) ELSE (
+    echo [+] Downloading RIFE frame-interpolation model...
+    git clone https://huggingface.co/TensorForger/RIFE-safetensors
+    IF ERRORLEVEL 1 (
+        echo [ERROR] Failed to clone RIFE model.
+        exit /b 1
+    )
+)
+
+:: ── FLUX.2-klein-4B base model ────────────────────────────────────────────────
+SET FLUX_DIR=FLUX.2-klein-4B
+SET FLUX_SENTINEL=FLUX.2-klein-4B\transformer\diffusion_pytorch_model.safetensors
+IF EXIST "%FLUX_SENTINEL%" (
+    echo [+] FLUX.2-klein-4B base model: already downloaded.
+) ELSE IF EXIST "%FLUX_DIR%\.git" (
+    echo [!] FLUX.2-klein-4B: directory exists but looks incomplete — resuming LFS download...
+    git -C "%FLUX_DIR%" pull --ff-only
+    git -C "%FLUX_DIR%" lfs pull
+) ELSE IF EXIST "%FLUX_DIR%" (
+    echo [!] FLUX.2-klein-4B: '%FLUX_DIR%' exists but is not a git repository.
+    echo [!]                  Remove it and re-run to download the model.
+) ELSE (
+    echo [+] Downloading FLUX.2-klein-4B base model...
+    git clone https://huggingface.co/black-forest-labs/FLUX.2-klein-4B
+    IF ERRORLEVEL 1 (
+        echo [ERROR] Failed to clone FLUX.2-klein-4B model.
+        exit /b 1
+    )
+)
+
+:: ── FLUX.2-klein-4B-int8 model ────────────────────────────────────────────────
+SET INT8_DIR=FLUX.2-klein-4B-int8
+SET INT8_SENTINEL=FLUX.2-klein-4B-int8\diffusion_pytorch_model.safetensors
+IF EXIST "%INT8_SENTINEL%" (
+    echo [+] FLUX.2-klein-4B int8 model: already downloaded.
+) ELSE IF EXIST "%INT8_DIR%\.git" (
+    echo [!] FLUX.2-klein-4B-int8: directory exists but looks incomplete — resuming LFS download...
+    git -C "%INT8_DIR%" pull --ff-only
+    git -C "%INT8_DIR%" lfs pull
+) ELSE IF EXIST "%INT8_DIR%" (
+    echo [!] FLUX.2-klein-4B-int8: '%INT8_DIR%' exists but is not a git repository.
+    echo [!]                       Remove it and re-run to download the model.
+) ELSE (
+    echo [+] Downloading FLUX.2-klein-4B int8 model...
+    git clone https://huggingface.co/aydin99/FLUX.2-klein-4B-int8
+    IF ERRORLEVEL 1 (
+        echo [ERROR] Failed to clone FLUX.2-klein-4B-int8 model.
+        exit /b 1
+    )
+)
+
+:: ── done ──────────────────────────────────────────────────────────────────────
+echo.
+echo [+] Installation complete.
+echo [!] Note: the GUI requires OBS to be installed for virtual webcam output.
+echo [!]       Download from https://obsproject.com/download
+echo.
+echo [+] Activate the environment and start:  .\%VENV_DIR%\Scripts\activate
+echo [+] Then run, for example:               python scripts\run_gradio_demo.py
+
+ENDLOCAL