Skip to content

Add Qwen3.5 0.8B and 2B variants #115

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Lokimorty wants to merge 6 commits into
base: main
Choose a base branch
from
Open

Add Qwen3.5 0.8B and 2B variants #115

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
638b3c6
Add Qwen3.5 0.8B and 2B model variants
Lokimorty Mar 3, 2026
6f54be0
Switch to mlx-swift-lm for Qwen3.5 support
Lokimorty Mar 3, 2026
29ebe28
Add MLXRandom dependency for clean builds
Lokimorty Mar 3, 2026
878f6b2
Fix App Shortcut phrase validation
Lokimorty Mar 3, 2026
73969fe
Add Qwen3.5 9B model option
Lokimorty Mar 3, 2026
760a827
Remove redundant direct Transformers dependency
Lokimorty Mar 3, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,9 @@
- [x] [Llama 3.2 1B](https://huggingface.co/mlx-community/Llama-3.2-1B-Instruct-4bit)
- [x] [Llama 3.2 3B](https://huggingface.co/mlx-community/Llama-3.2-3B-Instruct-4bit)
- [x] [DeepSeek-R1-Distill-Qwen-1.5B-4bit](https://huggingface.co/mlx-community/DeepSeek-R1-Distill-Qwen-1.5B-4bit)
- [x] [Qwen3.5-0.8B-4bit](https://huggingface.co/mlx-community/Qwen3.5-0.8B-4bit)
- [x] [Qwen3.5-2B-4bit](https://huggingface.co/mlx-community/Qwen3.5-2B-4bit)
- [x] [Qwen3.5-9B-4bit](https://huggingface.co/mlx-community/Qwen3.5-9B-4bit)

## credits
fullmoon is made possible by [MLX Swift](https://github.com/ml-explore/mlx-swift) from Apple. [MLX](https://github.com/ml-explore/mlx) is an array framework for machine learning research on Apple silicon.
36 changes: 18 additions & 18 deletions fullmoon.xcodeproj/project.pbxproj
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@
objects = {

/* Begin PBXBuildFile section */
25F645532D4E7A6F0049EFBE /* Transformers in Frameworks */ = {isa = PBXBuildFile; productRef = 25F645522D4E7A6F0049EFBE /* Transformers */; };
A1B2C3D42E9A100100AA55BB /* MLXRandom in Frameworks */ = {isa = PBXBuildFile; productRef = A1B2C3D52E9A100100AA55BB /* MLXRandom */; };
860F26A42CBC31D6004E8D40 /* MarkdownUI in Frameworks */ = {isa = PBXBuildFile; productRef = 860F26A32CBC31D6004E8D40 /* MarkdownUI */; };
869B97622D0DD46B0078DF5A /* MLXLMCommon in Frameworks */ = {isa = PBXBuildFile; productRef = 869B97612D0DD46B0078DF5A /* MLXLMCommon */; };
869B97642D0DD4D80078DF5A /* MLXLLM in Frameworks */ = {isa = PBXBuildFile; productRef = 869B97632D0DD4D80078DF5A /* MLXLLM */; };
Expand Down Expand Up @@ -45,7 +45,7 @@
files = (
869B97642D0DD4D80078DF5A /* MLXLLM in Frameworks */,
869B97622D0DD46B0078DF5A /* MLXLMCommon in Frameworks */,
25F645532D4E7A6F0049EFBE /* Transformers in Frameworks */,
A1B2C3D42E9A100100AA55BB /* MLXRandom in Frameworks */,
860F26A42CBC31D6004E8D40 /* MarkdownUI in Frameworks */,
);
runOnlyForDeploymentPostprocessing = 0;
Expand Down Expand Up @@ -100,7 +100,7 @@
860F26A32CBC31D6004E8D40 /* MarkdownUI */,
869B97612D0DD46B0078DF5A /* MLXLMCommon */,
869B97632D0DD4D80078DF5A /* MLXLLM */,
25F645522D4E7A6F0049EFBE /* Transformers */,
A1B2C3D52E9A100100AA55BB /* MLXRandom */,
);
productName = fullmoon;
productReference = 860E9CCE2CB055B000C5BB52 /* fullmoon.app */;
Expand Down Expand Up @@ -131,9 +131,9 @@
mainGroup = 860E9CC52CB055B000C5BB52;
minimizedProjectReferenceProxies = 1;
packageReferences = (
860E9CE22CB0564600C5BB52 /* XCRemoteSwiftPackageReference "mlx-swift-examples" */,
860E9CE22CB0564600C5BB52 /* XCRemoteSwiftPackageReference "mlx-swift-lm" */,
A1B2C3D62E9A100100AA55BB /* XCRemoteSwiftPackageReference "mlx-swift" */,
860F26A22CBC31D6004E8D40 /* XCRemoteSwiftPackageReference "swift-markdown-ui" */,
25F645512D4E7A6F0049EFBE /* XCRemoteSwiftPackageReference "swift-transformers" */,
);
preferredProjectObjectVersion = 77;
productRefGroup = 860E9CCF2CB055B000C5BB52 /* Products */;
Expand Down Expand Up @@ -396,17 +396,17 @@
/* End XCConfigurationList section */

/* Begin XCRemoteSwiftPackageReference section */
25F645512D4E7A6F0049EFBE /* XCRemoteSwiftPackageReference "swift-transformers" */ = {
A1B2C3D62E9A100100AA55BB /* XCRemoteSwiftPackageReference "mlx-swift" */ = {
isa = XCRemoteSwiftPackageReference;
repositoryURL = "https://github.com/huggingface/swift-transformers.git";
repositoryURL = "https://github.com/ml-explore/mlx-swift";
requirement = {
kind = upToNextMajorVersion;
minimumVersion = 0.1.17;
kind = upToNextMinorVersion;
minimumVersion = 0.30.6;
};
};
860E9CE22CB0564600C5BB52 /* XCRemoteSwiftPackageReference "mlx-swift-examples" */ = {
860E9CE22CB0564600C5BB52 /* XCRemoteSwiftPackageReference "mlx-swift-lm" */ = {
isa = XCRemoteSwiftPackageReference;
repositoryURL = "https://github.com/ml-explore/mlx-swift-examples/";
repositoryURL = "https://github.com/ml-explore/mlx-swift-lm";
requirement = {
branch = main;
kind = branch;
Expand All @@ -423,26 +423,26 @@
/* End XCRemoteSwiftPackageReference section */

/* Begin XCSwiftPackageProductDependency section */
25F645522D4E7A6F0049EFBE /* Transformers */ = {
isa = XCSwiftPackageProductDependency;
package = 25F645512D4E7A6F0049EFBE /* XCRemoteSwiftPackageReference "swift-transformers" */;
productName = Transformers;
};
860F26A32CBC31D6004E8D40 /* MarkdownUI */ = {
isa = XCSwiftPackageProductDependency;
package = 860F26A22CBC31D6004E8D40 /* XCRemoteSwiftPackageReference "swift-markdown-ui" */;
productName = MarkdownUI;
};
869B97612D0DD46B0078DF5A /* MLXLMCommon */ = {
isa = XCSwiftPackageProductDependency;
package = 860E9CE22CB0564600C5BB52 /* XCRemoteSwiftPackageReference "mlx-swift-examples" */;
package = 860E9CE22CB0564600C5BB52 /* XCRemoteSwiftPackageReference "mlx-swift-lm" */;
productName = MLXLMCommon;
};
869B97632D0DD4D80078DF5A /* MLXLLM */ = {
isa = XCSwiftPackageProductDependency;
package = 860E9CE22CB0564600C5BB52 /* XCRemoteSwiftPackageReference "mlx-swift-examples" */;
package = 860E9CE22CB0564600C5BB52 /* XCRemoteSwiftPackageReference "mlx-swift-lm" */;
productName = MLXLLM;
};
A1B2C3D52E9A100100AA55BB /* MLXRandom */ = {
isa = XCSwiftPackageProductDependency;
package = A1B2C3D62E9A100100AA55BB /* XCRemoteSwiftPackageReference "mlx-swift" */;
productName = MLXRandom;
};
/* End XCSwiftPackageProductDependency section */
};
rootObject = 860E9CC62CB055B000C5BB52 /* Project object */;
Expand Down
21 changes: 21 additions & 0 deletions fullmoon/Models/Models.swift
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,9 @@ public extension ModelConfiguration {
switch self {
case .deepseek_r1_distill_qwen_1_5b_4bit: .reasoning
case .deepseek_r1_distill_qwen_1_5b_8bit: .reasoning
case .qwen_3_5_0_8b_4bit: .reasoning
case .qwen_3_5_2b_4bit: .reasoning
case .qwen_3_5_9b_4bit: .reasoning
case .qwen_3_4b_4bit: .reasoning
case .qwen_3_8b_4bit: .reasoning
default: .regular
Expand Down Expand Up @@ -45,6 +48,18 @@ extension ModelConfiguration: @retroactive Equatable {
id: "mlx-community/DeepSeek-R1-Distill-Qwen-1.5B-8bit"
)

public static let qwen_3_5_0_8b_4bit = ModelConfiguration(
id: "mlx-community/Qwen3.5-0.8B-4bit"
)

public static let qwen_3_5_2b_4bit = ModelConfiguration(
id: "mlx-community/Qwen3.5-2B-4bit"
)

public static let qwen_3_5_9b_4bit = ModelConfiguration(
id: "mlx-community/Qwen3.5-9B-4bit"
)

public static let qwen_3_4b_4bit = ModelConfiguration(
id: "mlx-community/Qwen3-4B-4bit"
)
Expand All @@ -58,6 +73,9 @@ extension ModelConfiguration: @retroactive Equatable {
llama_3_2_3b_4bit,
deepseek_r1_distill_qwen_1_5b_4bit,
deepseek_r1_distill_qwen_1_5b_8bit,
qwen_3_5_0_8b_4bit,
qwen_3_5_2b_4bit,
qwen_3_5_9b_4bit,
qwen_3_4b_4bit,
qwen_3_8b_4bit,
]
Expand Down Expand Up @@ -118,6 +136,9 @@ extension ModelConfiguration: @retroactive Equatable {
case .llama_3_2_3b_4bit: return 1.8
case .deepseek_r1_distill_qwen_1_5b_4bit: return 1.0
case .deepseek_r1_distill_qwen_1_5b_8bit: return 1.9
case .qwen_3_5_0_8b_4bit: return 0.7
case .qwen_3_5_2b_4bit: return 1.8
case .qwen_3_5_9b_4bit: return 5.5
case .qwen_3_4b_4bit: return 2.3
case .qwen_3_8b_4bit: return 4.7
default: return nil
Expand Down
2 changes: 1 addition & 1 deletion fullmoon/Models/RequestLLMIntent.swift
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -91,7 +91,7 @@ struct NewChatShortcut: AppShortcutsProvider {
AppShortcut(
intent: RequestLLMIntent(),
phrases: [
"Start a new chat",
"Start a new \(.applicationName) chat",
"Start a \(.applicationName) chat",
"Chat with \(.applicationName)",
"Ask \(.applicationName) a question"
Expand Down