# Qwen3.8-Flash-Next GSQ-RCO Coder (Pruned GGUF) lokal ausprobieren – Windows (PowerShell)
# Automatisch erzeugt von Tokenstrom (tokenstrom.com).
# Aufruf:  powershell -ExecutionPolicy Bypass -File try_windows.ps1

$ErrorActionPreference = "Stop"
Set-Location $PSScriptRoot
Write-Host "== Qwen3.8-Flash-Next GSQ-RCO Coder (Pruned GGUF) =="
if (Get-Command nvidia-smi -ErrorAction SilentlyContinue) { nvidia-smi --query-gpu=name,memory.total --format=csv,noheader }
Write-Host "WARNUNG: Groß (115,9 B parameters, MoE): running entirely on the GPU takes a lot of VRAM (video memory). With MoE offloading, a regular graphics card plus plenty of RAM is enough (see table). Colab Free isn't enough." -ForegroundColor Yellow

if ($true) {
  Write-Host "== Ollama =="
  if (-not (Get-Command ollama -ErrorAction SilentlyContinue)) {
    Write-Host "Installing Ollama ..."
    winget install -e --id Ollama.Ollama --accept-source-agreements --accept-package-agreements
    $env:Path = [Environment]::GetEnvironmentVariable("Path","Machine") + ";" + [Environment]::GetEnvironmentVariable("Path","User")
  }
  try { Invoke-WebRequest -UseBasicParsing http://localhost:11434/api/tags | Out-Null } catch { Start-Process -WindowStyle Hidden ollama -ArgumentList "serve"; Start-Sleep 3 }
  if (-not $env:MODEL) { $env:MODEL = "hf.co/Skyrion/Qwen3.8-Flash-Next-GSQ-RCO-Coder-GGUF-Pruned" }
  Write-Host "Downloading $($env:MODEL) (the first time this takes a few minutes, depending on size) ..."
  ollama pull $env:MODEL
  Write-Host ""
  Write-Host "Done! Now type your question and press Enter. Quit with /bye" -ForegroundColor Green
  ollama run $env:MODEL
  exit
}