feat: migration to mlx-swift-lm v3

MLXServerTests/Server/ModelBackedInferenceValidationTests.swift

@@ -1,7 +1,9 @@
 import Foundation
-import Hub
+import HuggingFace
+import MLXHuggingFace
 import MLXLMCommon
 import MLXVLM
+import Tokenizers
 import XCTest
 @testable import MLX_Server
 
@@ -671,10 +673,9 @@ private actor LocalGemmaFixture {
         }
 
         let loadTask = Task<ModelContainer, Error> {
-            let cachesDir = FileManager.default.urls(for: .cachesDirectory, in: .userDomainMask).first
-            let hub = HubApi(downloadBase: cachesDir, cache: nil)
             return try await VLMModelFactory.shared.loadContainer(
-                hub: hub,
+                from: #hubDownloader(HubClient.default),
+                using: #huggingFaceTokenizerLoader(),
                 configuration: ModelConfiguration(directory: localDir),
                 progressHandler: { _ in }
             )