ollama: cleanup docs

modules/services/ollama.nix

@@ -9,6 +9,9 @@
 # but the service will need modification to allow net access first.
 #
 # models are defined e.g. here: <https://ollama.com/library/llama3.2:3b/blobs/dde5aa3fc5ff>
+#
+# TODO: enable ShellGPT integration?
+# - <https://github.com/TheR1D/shell_gpt?tab=readme-ov-file#docker--ollama>
 { config, lib, pkgs, ... }:
 let
   cfg = config.sane.services.ollama;
@@ -26,9 +29,7 @@ in
     services.ollama.models = models;
 
     # these acceleration settings are relevant to `desko`.
-    # docs also suggest nixpkgs needs to be built with `rocmSupport` (which it isn't, by default),
-    # but i still see GPU offload even without that.
-    services.ollama.acceleration = "rocm";  # AMD GPU acceleration
+    services.ollama.acceleration = "rocm";  # AMD GPU acceleration (achieves the same as `nixpkgs.config.rocmSupport = true` but just for ollama)
     services.ollama.rocmOverrideGfx = "10.1.0";  #< `nix-shell -p "rocmPackages.rocminfo" --run "rocminfo" | grep "gfx"`
     # services.ollama.environmentVariables.HCC_AMDGPU_TARGET = "gfx1010";  # seems to be unnecessary