chore(example/llm): minor refactor (#227)

juspay · Jun 14, 2024 · 8010b5a · 8010b5a
1 parent db7ab71
commit 8010b5a
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 8 deletions.
diff --git a/example/llm/flake.lock b/example/llm/flake.lock
diff --git a/example/llm/flake.nix b/example/llm/flake.nix
@@ -21,9 +21,11 @@
           # Backend service to perform inference on LLM models
           ollama."ollama1" = {
             enable = true;
-            # The models are usually huge, downloading them in every project directory can lead to a lot of duplication
-            dataDir = "$HOME/.services-flake/ollama1";
-            models = [ "llama2-uncensored" ];
+            # The models are usually huge, downloading them in every project
+            # directory can lead to a lot of duplication. Change here to a
+            # directory where the Ollama models can be stored and shared across
+            # projects.
+            # dataDir = "$HOME/.services-flake/ollama1";
           };
           # Get ChatGPT like UI, but open-source, with Open WebUI
           open-webui."open-webui1" = {
@@ -33,9 +35,15 @@
                 inherit (pc.config.services.ollama.ollama1) host port;
               in
               {
-                OLLAMA_API_BASE_URL = "http://${host}:${toString port}";
+                ENABLE_OLLAMA_API = "True";
+                OLLAMA_BASE_URL = "http://${host}:${toString port}";
+                OLLAMA_API_BASE_URL = "http://${host}:${toString port}/api";
                 WEBUI_AUTH = "False";
-              };
+                DEVICE_TYPE = "cpu";
+                RAG_EMBEDDING_ENGINE = "ollama";
+                RAG_EMBEDDING_MODEL = "mxbai-embed-large:latest";
+                RAG_EMBEDDING_MODEL_AUTO_UPDATE = "True";
+                RAG_RERANKING_MODEL_AUTO_UPDATE = "True";              };
           };
         };