Show pip progress for PyTorch download on Windows

The torch CUDA wheel is ~2.8 GB and the CPU wheel is ~300 MB. With
| Out-Null suppressing all output, the install appeared completely
frozen with no feedback. Remove | Out-Null for the torch install
lines so pip's download progress bar is visible. Add a size hint
so users know the download is expected to take a while.

Also moves the Triton success message inside the GPU branch so it
only prints when Triton was actually installed.
This commit is contained in:
Daniel Han 2026-03-18 05:33:10 +00:00
commit a7d4e8860f

View file

@ -886,17 +886,16 @@ Write-Host "[OK] TORCHINDUCTOR_CACHE_DIR set to $TorchCacheDir (avoids MAX_PATH
if ($HasNvidiaSmi) {
$CuTag = Get-PytorchCudaTag
Write-Host " Installing PyTorch with CUDA support ($CuTag)..." -ForegroundColor Cyan
pip install torch torchvision torchaudio --index-url "https://download.pytorch.org/whl/$CuTag" 2>&1 | Out-Null
Write-Host " (This download is ~2.8 GB -- may take a few minutes)" -ForegroundColor Gray
pip install torch torchvision torchaudio --index-url "https://download.pytorch.org/whl/$CuTag"
# Install Triton for Windows (enables torch.compile -- without it training can hang)
Write-Host " Installing Triton for Windows..." -ForegroundColor Cyan
pip install "triton-windows<3.7" 2>&1 | Out-Null
Write-Host "[OK] Triton for Windows installed (enables torch.compile)" -ForegroundColor Green
} else {
Write-Host " Installing PyTorch (CPU-only)..." -ForegroundColor Cyan
pip install torch torchvision torchaudio 2>&1 | Out-Null
}
if ($HasNvidiaSmi) {
Write-Host "[OK] Triton for Windows installed (enables torch.compile)" -ForegroundColor Green
pip install torch torchvision torchaudio
}
# Ordered heavy dependency installation — shared cross-platform script