fix: idols-ai - disable zram swap when running large LLM models via mmap with limited RAM

hosts/idols-ai/default.nix

@@ -24,6 +24,9 @@ in
     ./secureboot.nix
   ];
 
+  # Zram consumes physical memory for compression, which can cause a deadlock and system hang if the model size approaches the physical memory limit.
+  zramSwap.enable = lib.mkForce false;
+
   services.sunshine.enable = lib.mkForce true;
   services.tuned.ppdSettings.main.default = lib.mkForce "performance";