Add Qwen3-1.7B iOS and macOS export support

Package.swift

@@ -23,6 +23,12 @@ let package = Package(
                 "CoreAIDiffusionPipeline"
             ]
         ),
+        .library(
+            name: "CoreAIVideoDiffusion",
+            targets: [
+                "CoreAIVideoDiffusionPipeline"
+            ]
+        ),
         .library(
             name: "CoreAISegmentation",
             targets: [
@@ -44,6 +50,7 @@ let package = Package(
         .package(url: "https://github.com/apple/swift-argument-parser", from: "1.2.0"),
         .package(url: "https://github.com/huggingface/swift-transformers", from: "1.1.0"),
         .package(url: "https://github.com/mlc-ai/xgrammar", exact: "0.2.2"),
+        .package(url: "https://github.com/hummingbird-project/hummingbird", exact: "2.22.0"),
     ],
     targets: [
         .target(
@@ -116,6 +123,29 @@ let package = Package(
             ]
         ),
 
+        .target(
+            name: "CoreAIVideoDiffusionPipeline",
+            dependencies: [
+                "CoreAIDiffusionPipeline",
+                "CoreAIShared",
+                .product(name: "Transformers", package: "swift-transformers"),
+            ],
+            path: "swift/Sources/CoreAIVideoDiffusionPipeline",
+            swiftSettings: [
+                .enableUpcomingFeature("MemberImportVisibility")
+            ]
+        ),
+
+        // Shared types for LLM CLI tools (used by both llm-runner and llm-server)
+        .target(
+            name: "CoreAILMCommon",
+            dependencies: [],
+            path: "swift/Sources/CoreAILMCommon",
+            swiftSettings: [
+                .enableUpcomingFeature("MemberImportVisibility")
+            ]
+        ),
+
         // CXGrammar C bridge
         .target(
             name: "CXGrammar",
@@ -140,6 +170,20 @@ let package = Package(
                 .enableUpcomingFeature("MemberImportVisibility")
             ]
         ),
+        .executableTarget(
+            name: "llm-server",
+            dependencies: [
+                "CoreAILanguageModels",
+                "CoreAILMCommon",
+                "CoreAIShared",
+                .product(name: "ArgumentParser", package: "swift-argument-parser"),
+                .product(name: "Hummingbird", package: "hummingbird"),
+            ],
+            path: "swift/Sources/Tools/llm-server",
+            swiftSettings: [
+                .enableUpcomingFeature("MemberImportVisibility")
+            ]
+        ),
         .executableTarget(
             name: "image-segmenter",
             dependencies: [
@@ -176,6 +220,18 @@ let package = Package(
                 .enableUpcomingFeature("MemberImportVisibility")
             ]
         ),
+        .executableTarget(
+            name: "videodiffusion-runner",
+            dependencies: [
+                "CoreAIVideoDiffusionPipeline",
+                "CoreAIShared",
+                .product(name: "ArgumentParser", package: "swift-argument-parser"),
+            ],
+            path: "swift/Sources/Tools/videodiffusion-runner",
+            swiftSettings: [
+                .enableUpcomingFeature("MemberImportVisibility")
+            ]
+        ),
         .executableTarget(
             name: "speech-recognizer",
             dependencies: [
@@ -245,6 +301,7 @@ let package = Package(
             name: "DiffusionPipelineTests",
             dependencies: [
                 "CoreAIDiffusionPipeline",
+                "CoreAIVideoDiffusionPipeline",
                 "TestUtilities",
             ],
             path: "swift/Tests/DiffusionPipelineTests"
@@ -259,6 +316,11 @@ let package = Package(
             dependencies: ["CoreAIShared", "TestUtilities"],
             path: "swift/Tests/CoreAISharedTests"
         ),
+        .testTarget(
+            name: "CoreAILMCommonTests",
+            dependencies: ["CoreAILMCommon"],
+            path: "swift/Tests/CoreAILMCommonTests"
+        ),
         .testTarget(
             name: "GuidedGenerationTests",
             dependencies: [

models/README.md

@@ -60,6 +60,8 @@ uv run coreai.llm.export Qwen/Qwen3-0.6B --compression none
 uv run coreai.llm.export Qwen/Qwen3-0.6B --platform iOS --compression 4bit_weight_palettized_group8
 ```
 
+**Note:** By default, all quantization presets use `coreai-opt`'s `eager` execution mode. Use the `--quantization-mode graph` argument to override and use graph-mode quantization.
+
 ##### Specifying Compression Configs via YAML files
 
 Specialized compression recipes that aren't covered by pre-defined presets can be specified as YAML files using the `--compression-config` option with the path to a [coreai-opt](https://github.com/apple/coreai-optimization) config.
@@ -171,6 +173,8 @@ uv run models/<name>/export.py --include-debug-info   # embed debug information
 - [GPT-OSS](gpt_oss)
 - [Mistral](mistral)
 - [Mixtral](mixtral)
+- [Muse Glimmer](muse_glimmer)
+- [Phi](phi)
 - [Qwen2.5](qwen2)
 - [Qwen3](qwen3)
 - [Qwen3 MoE](qwen3_moe)