feat: Improve ollama2 model fetching

This commit introduces a more efficient approach to fetching ollama2 models, ensuring proper filtering and handling of providers. This enhances the robustness and reliability of the model loading process, streamlining the overall user experience.

src/db/models.ts

@@ -407,6 +407,10 @@ export const ollamaFormatAllCustomModels = async (
     (provider) => provider.provider === "llamafile"
   )
 
+  const ollamaProviders = allProviders.filter(
+    (provider) => provider.provider === "ollama2"
+  )
+
   const lmModelsPromises = lmstudioProviders.map((provider) =>
     dynamicFetchLMStudio({
       baseUrl: provider.baseUrl,
@@ -421,12 +425,11 @@ export const ollamaFormatAllCustomModels = async (
     })
   )
 
-  const ollamaModelsPromises = allProviders.map((provider) => (
+  const ollamaModelsPromises = ollamaProviders.map((provider) =>
     dynamicFetchOllama2({
       baseUrl: provider.baseUrl,
       providerId: provider.id
-    })
-  ))
+    }))
 
   const lmModelsFetch = await Promise.all(lmModelsPromises)