Update instructions for 7900XTX AMD Preview Driver 25.20.01.17 under Windows 11 (bare, no extra translation layer)

### Feature Idea
Update documentation for stack:

WINDOWS 11 - 7900XTX - ROCm

I rebuilt ComfyUI ROCm on bare windows using the AMD preview driver, and it worked pretty easily, so far I run Flux, Zimage and tested the Flux VAE decode up to 2048 pixel and went successfully with Hunyuan 3D 2.1 STL generation and it all works so far.

More testing is needed, I was warned of potential memory leaks on repeated inference, but I personally haven't encountered it yet. I have yet to try audio workflows and video workflows.

Running it bare with the new driver is so much easier than Zluda or WSL2.

Consider updating the instructions for people running AMD 7900XTX ([compatibility matrix](https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/compatibility/compatibilityrad/windows/windows_compatibility.html))

Instruction are surprisingly painless this time around thanks to no translation layer being needed

1) [Install the preview driver as per AMD instructions
](https://www.amd.com/en/resources/support-articles/release-notes/RN-AMDGPU-WINDOWS-PYTORCH-7-1-1.html)

2) Install ComfyUI with AMD pytorch dependencies, I made a powershell script meant to run on an empty folder, where it makes the UV local interpreter, and downloads the 13GB worth of python and AMD pytorch binaries

```ps1
# Ensure uv is installed
if (-not (Get-Command uv -ErrorAction SilentlyContinue)) {
Write-Host "Installing uv..."
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/0.9.16/install.ps1 | iex"
} else {
Write-Host "uv already installed."
}

# Clone ComfyUI into current directory
git init
git remote add origin https://github.com/comfyanonymous/ComfyUI.git
git fetch
git checkout -t origin/master

# Create venv with Python 3.12
uv venv .venv --python 3.12
.venv\Scripts\Activate.ps1

# Install ROCm SDK and libraries
uv pip install --no-cache-dir `
https://repo.radeon.com/rocm/windows/rocm-rel-7.1.1/rocm_sdk_core-0.1.dev0-py3-none-win_amd64.whl `
https://repo.radeon.com/rocm/windows/rocm-rel-7.1.1/rocm_sdk_devel-0.1.dev0-py3-none-win_amd64.whl `
https://repo.radeon.com/rocm/windows/rocm-rel-7.1.1/rocm_sdk_libraries_custom-0.1.dev0-py3-none-win_amd64.whl `
https://repo.radeon.com/rocm/windows/rocm-rel-7.1.1/rocm-0.1.dev0.tar.gz

# Install ROCm-enabled PyTorch stack
uv pip install --no-cache-dir `
https://repo.radeon.com/rocm/windows/rocm-rel-7.1.1/torch-2.9.0+rocmsdk20251116-cp312-cp312-win_amd64.whl `
https://repo.radeon.com/rocm/windows/rocm-rel-7.1.1/torchaudio-2.9.0+rocmsdk20251116-cp312-cp312-win_amd64.whl `
https://repo.radeon.com/rocm/windows/rocm-rel-7.1.1/torchvision-0.24.0+rocmsdk20251116-cp312-cp312-win_amd64.whl

# Verify installation
try {
python -c "import torch" | Out-Null
Write-Host "Torch import: Success"
} catch {
Write-Host "Torch import: Failure"
}

Write-Host "CUDA available:" (python -c "import torch; print(torch.cuda.is_available())")
Write-Host "Device name:" (python -c "import torch; print(torch.cuda.get_device_name(0))")
python -m torch.utils.collect_env

# Install ComfyUI custom node manager
Set-Location custom_nodes
git clone https://github.com/ltdrdata/ComfyUI-Manager comfyui-manager
Set-Location ..

# Install ComfyUI requirements
uv pip install -r "requirements.txt"

# Run ComfyUI
uv run main.py
```

[LOG ](https://github.com/OrsoEric/HOWTO-ComfyUI/blob/Master/logs/2025-12-08%20Trying%20ROCm%20Windows.md)

[README IN PROGRESS](https://github.com/OrsoEric/HOWTO-ComfyUI/tree/Master)

Start by comparing the existing installation documentation with the linked HOWTO log and README IN PROGRESS; the issue does not identify a ComfyUI documentation file. Use the supplied PowerShell setup, requirements.txt, and main.py entry point as references, and consider the work done when bare Windows 11 7900XTX ROCm setup instructions are documented clearly.

Written by the indexing model from the issue text.