The Swift Package Index logo.Swift Package Index

Has it really been five years since Swift Package Index launched? Read our anniversary blog post!

Build Information

Successful build of mlx-swift-lm, reference main (74f85d), with Swift 6.2 for macOS (SPM) on 9 Dec 2025 01:30:28 UTC.

Swift 6 data race errors: 0

Build Command

env DEVELOPER_DIR=/Applications/Xcode-26.0.0.app xcrun swift build --arch arm64

Build Log

/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:70:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:71:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
 73 |         ]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:72:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 73 |         ]
 74 |     }
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/AfMoE.swift:502:22: warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
500 |         // Create attention masks
501 |         let faCache: [KVCache]? = layerCache.map { [$0[faIdx]] }
502 |         let faMask = createAttentionMask(h: h, cache: faCache)
    |                      `- warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
503 |
504 |         var swaMask: MLXFast.ScaledDotProductAttentionMaskMode = .none
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:525:17: warning: returning task-isolated 'model' as a 'sending' result risks causing data races; this is an error in the Swift 6 language mode
523 |             messageGenerator: messageGenerator)
524 |
525 |         return .init(
    |                 |- warning: returning task-isolated 'model' as a 'sending' result risks causing data races; this is an error in the Swift 6 language mode
    |                 `- note: returning task-isolated 'model' risks causing data races since the caller assumes that 'model' can be safely sent to other isolation domains
526 |             configuration: configuration, model: model, processor: processor, tokenizer: tokenizer)
527 |     }
[#DeprecatedDeclaration]: <https://docs.swift.org/compiler/documentation/diagnostics/deprecated-declaration>
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModel.swift:27:13: warning: variable 'state' was never mutated; consider changing to 'let' constant
25 |         let prefillStepSize = windowSize ?? 512
26 |         var y = input.text
27 |         var state: LMOutput.State? = nil
   |             `- warning: variable 'state' was never mutated; consider changing to 'let' constant
28 |
29 |         // prepare the prompt in chunks if larger than the prefill size
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModel.swift:32:17: warning: initialization of immutable value 'result' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
30 |         while y.tokens.size > prefillStepSize {
31 |             let input = y[.newAxis, ..<prefillStepSize]
32 |             let result = self(input, cache: cache.isEmpty ? nil : cache, state: state)
   |                 `- warning: initialization of immutable value 'result' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
33 |             eval(cache)
34 |             y = y[prefillStepSize...]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:461:14: warning: non-final class 'LLMModelFactory' cannot conform to 'Sendable'; use '@unchecked Sendable'; this is an error in the Swift 6 language mode
459 | ///     configuration: LLMRegistry.llama3_8B_4bit)
460 | /// ```
461 | public class LLMModelFactory: ModelFactory {
    |              `- warning: non-final class 'LLMModelFactory' cannot conform to 'Sendable'; use '@unchecked Sendable'; this is an error in the Swift 6 language mode
462 |
463 |     public init(typeRegistry: ModelTypeRegistry, modelRegistry: AbstractModelRegistry) {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:31:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 29 |     private static func all() -> [String: @Sendable (URL) throws -> any LanguageModel] {
 30 |         [
 31 |             "mistral": create(LlamaConfiguration.self, LlamaModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:32:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 30 |         [
 31 |             "mistral": create(LlamaConfiguration.self, LlamaModel.init),
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:33:20: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 31 |             "mistral": create(LlamaConfiguration.self, LlamaModel.init),
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
    |                    `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:34:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:35:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:36:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:37:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:38:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:39:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:40:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:41:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:42:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:43:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:44:27: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
    |                           `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:45:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:46:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:47:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:48:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
 50 |             "granitemoehybrid": create(
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:49:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 50 |             "granitemoehybrid": create(
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:50:33: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
 50 |             "granitemoehybrid": create(
    |                                 `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:52:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 50 |             "granitemoehybrid": create(
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:53:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:54:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:55:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:56:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:57:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:58:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:59:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:60:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:61:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:62:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:63:27: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
    |                           `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:64:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:65:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:66:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:67:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:68:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:69:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:70:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:71:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
 73 |         ]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:72:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 73 |         ]
 74 |     }
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/AfMoE.swift:502:22: warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
500 |         // Create attention masks
501 |         let faCache: [KVCache]? = layerCache.map { [$0[faIdx]] }
502 |         let faMask = createAttentionMask(h: h, cache: faCache)
    |                      `- warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
503 |
504 |         var swaMask: MLXFast.ScaledDotProductAttentionMaskMode = .none
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:525:17: warning: returning task-isolated 'model' as a 'sending' result risks causing data races; this is an error in the Swift 6 language mode
523 |             messageGenerator: messageGenerator)
524 |
525 |         return .init(
    |                 |- warning: returning task-isolated 'model' as a 'sending' result risks causing data races; this is an error in the Swift 6 language mode
    |                 `- note: returning task-isolated 'model' risks causing data races since the caller assumes that 'model' can be safely sent to other isolation domains
526 |             configuration: configuration, model: model, processor: processor, tokenizer: tokenizer)
527 |     }
[#DeprecatedDeclaration]: <https://docs.swift.org/compiler/documentation/diagnostics/deprecated-declaration>
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModel.swift:27:13: warning: variable 'state' was never mutated; consider changing to 'let' constant
25 |         let prefillStepSize = windowSize ?? 512
26 |         var y = input.text
27 |         var state: LMOutput.State? = nil
   |             `- warning: variable 'state' was never mutated; consider changing to 'let' constant
28 |
29 |         // prepare the prompt in chunks if larger than the prefill size
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModel.swift:32:17: warning: initialization of immutable value 'result' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
30 |         while y.tokens.size > prefillStepSize {
31 |             let input = y[.newAxis, ..<prefillStepSize]
32 |             let result = self(input, cache: cache.isEmpty ? nil : cache, state: state)
   |                 `- warning: initialization of immutable value 'result' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
33 |             eval(cache)
34 |             y = y[prefillStepSize...]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:461:14: warning: non-final class 'LLMModelFactory' cannot conform to 'Sendable'; use '@unchecked Sendable'; this is an error in the Swift 6 language mode
459 | ///     configuration: LLMRegistry.llama3_8B_4bit)
460 | /// ```
461 | public class LLMModelFactory: ModelFactory {
    |              `- warning: non-final class 'LLMModelFactory' cannot conform to 'Sendable'; use '@unchecked Sendable'; this is an error in the Swift 6 language mode
462 |
463 |     public init(typeRegistry: ModelTypeRegistry, modelRegistry: AbstractModelRegistry) {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:31:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 29 |     private static func all() -> [String: @Sendable (URL) throws -> any LanguageModel] {
 30 |         [
 31 |             "mistral": create(LlamaConfiguration.self, LlamaModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:32:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 30 |         [
 31 |             "mistral": create(LlamaConfiguration.self, LlamaModel.init),
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:33:20: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 31 |             "mistral": create(LlamaConfiguration.self, LlamaModel.init),
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
    |                    `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:34:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:35:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:36:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:37:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:38:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:39:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:40:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:41:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:42:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:43:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:44:27: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
    |                           `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:45:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:46:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:47:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:48:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
 50 |             "granitemoehybrid": create(
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:49:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 50 |             "granitemoehybrid": create(
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:50:33: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
 50 |             "granitemoehybrid": create(
    |                                 `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:52:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 50 |             "granitemoehybrid": create(
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:53:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:54:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:55:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:56:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:57:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:58:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:59:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:60:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:61:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:62:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:63:27: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
    |                           `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:64:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:65:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:66:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:67:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:68:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:69:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:70:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:71:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
 73 |         ]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:72:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 73 |         ]
 74 |     }
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/AfMoE.swift:502:22: warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
500 |         // Create attention masks
501 |         let faCache: [KVCache]? = layerCache.map { [$0[faIdx]] }
502 |         let faMask = createAttentionMask(h: h, cache: faCache)
    |                      `- warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
503 |
504 |         var swaMask: MLXFast.ScaledDotProductAttentionMaskMode = .none
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:525:17: warning: returning task-isolated 'model' as a 'sending' result risks causing data races; this is an error in the Swift 6 language mode
523 |             messageGenerator: messageGenerator)
524 |
525 |         return .init(
    |                 |- warning: returning task-isolated 'model' as a 'sending' result risks causing data races; this is an error in the Swift 6 language mode
    |                 `- note: returning task-isolated 'model' risks causing data races since the caller assumes that 'model' can be safely sent to other isolation domains
526 |             configuration: configuration, model: model, processor: processor, tokenizer: tokenizer)
527 |     }
[#DeprecatedDeclaration]: <https://docs.swift.org/compiler/documentation/diagnostics/deprecated-declaration>
[492/512] Compiling MLXLLM AfMoE.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModel.swift:27:13: warning: variable 'state' was never mutated; consider changing to 'let' constant
25 |         let prefillStepSize = windowSize ?? 512
26 |         var y = input.text
27 |         var state: LMOutput.State? = nil
   |             `- warning: variable 'state' was never mutated; consider changing to 'let' constant
28 |
29 |         // prepare the prompt in chunks if larger than the prefill size
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModel.swift:32:17: warning: initialization of immutable value 'result' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
30 |         while y.tokens.size > prefillStepSize {
31 |             let input = y[.newAxis, ..<prefillStepSize]
32 |             let result = self(input, cache: cache.isEmpty ? nil : cache, state: state)
   |                 `- warning: initialization of immutable value 'result' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
33 |             eval(cache)
34 |             y = y[prefillStepSize...]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:461:14: warning: non-final class 'LLMModelFactory' cannot conform to 'Sendable'; use '@unchecked Sendable'; this is an error in the Swift 6 language mode
459 | ///     configuration: LLMRegistry.llama3_8B_4bit)
460 | /// ```
461 | public class LLMModelFactory: ModelFactory {
    |              `- warning: non-final class 'LLMModelFactory' cannot conform to 'Sendable'; use '@unchecked Sendable'; this is an error in the Swift 6 language mode
462 |
463 |     public init(typeRegistry: ModelTypeRegistry, modelRegistry: AbstractModelRegistry) {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:31:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 29 |     private static func all() -> [String: @Sendable (URL) throws -> any LanguageModel] {
 30 |         [
 31 |             "mistral": create(LlamaConfiguration.self, LlamaModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:32:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 30 |         [
 31 |             "mistral": create(LlamaConfiguration.self, LlamaModel.init),
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:33:20: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 31 |             "mistral": create(LlamaConfiguration.self, LlamaModel.init),
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
    |                    `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:34:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 32 |             "llama": create(LlamaConfiguration.self, LlamaModel.init),
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:35:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 33 |             "phi": create(PhiConfiguration.self, PhiModel.init),
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:36:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 34 |             "phi3": create(Phi3Configuration.self, Phi3Model.init),
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:37:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 35 |             "phimoe": create(PhiMoEConfiguration.self, PhiMoEModel.init),
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:38:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 36 |             "gemma": create(GemmaConfiguration.self, GemmaModel.init),
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:39:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 37 |             "gemma2": create(Gemma2Configuration.self, Gemma2Model.init),
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:40:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 38 |             "gemma3": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:41:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 39 |             "gemma3_text": create(Gemma3TextConfiguration.self, Gemma3TextModel.init),
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:42:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 40 |             "gemma3n": create(Gemma3nTextConfiguration.self, Gemma3nTextModel.init),
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:43:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 41 |             "qwen2": create(Qwen2Configuration.self, Qwen2Model.init),
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:44:27: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 42 |             "qwen3": create(Qwen3Configuration.self, Qwen3Model.init),
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
    |                           `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:45:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 43 |             "qwen3_moe": create(Qwen3MoEConfiguration.self, Qwen3MoEModel.init),
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:46:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 44 |             "starcoder2": create(Starcoder2Configuration.self, Starcoder2Model.init),
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:47:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 45 |             "cohere": create(CohereConfiguration.self, CohereModel.init),
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:48:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 46 |             "openelm": create(OpenElmConfiguration.self, OpenELMModel.init),
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
 50 |             "granitemoehybrid": create(
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:49:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 47 |             "internlm2": create(InternLM2Configuration.self, InternLM2Model.init),
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 50 |             "granitemoehybrid": create(
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:50:33: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 48 |             "deepseek_v3": create(DeepseekV3Configuration.self, DeepseekV3Model.init),
 49 |             "granite": create(GraniteConfiguration.self, GraniteModel.init),
 50 |             "granitemoehybrid": create(
    |                                 `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:52:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 50 |             "granitemoehybrid": create(
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:53:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 51 |                 GraniteMoeHybridConfiguration.self, GraniteMoeHybridModel.init),
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:54:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 52 |             "mimo": create(MiMoConfiguration.self, MiMoModel.init),
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:55:26: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 53 |             "glm4": create(GLM4Configuration.self, GLM4Model.init),
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
    |                          `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:56:23: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 54 |             "acereason": create(Qwen2Configuration.self, Qwen2Model.init),
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
    |                       `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:57:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 55 |             "falcon_h1": create(FalconH1Configuration.self, FalconH1Model.init),
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:58:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 56 |             "bitnet": create(BitnetConfiguration.self, BitnetModel.init),
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:59:21: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 57 |             "smollm3": create(SmolLM3Configuration.self, SmolLM3Model.init),
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
    |                     `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:60:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 58 |             "ernie4_5": create(Ernie45Configuration.self, Ernie45Model.init),
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:61:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 59 |             "lfm2": create(LFM2Configuration.self, LFM2Model.init),
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:62:24: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 60 |             "baichuan_m1": create(BaichuanM1Configuration.self, BaichuanM1Model.init),
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
    |                        `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:63:27: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 61 |             "exaone4": create(Exaone4Configuration.self, Exaone4Model.init),
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
    |                           `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:64:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 62 |             "gpt_oss": create(GPTOSSConfiguration.self, GPTOSSModel.init),
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:65:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 63 |             "lille-130m": create(Lille130mConfiguration.self, Lille130mModel.init),
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:66:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 64 |             "olmoe": create(OlmoEConfiguration.self, OlmoEModel.init),
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:67:28: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 65 |             "olmo2": create(Olmo2Configuration.self, Olmo2Model.init),
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
    |                            `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:68:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 66 |             "olmo3": create(Olmo3Configuration.self, Olmo3Model.init),
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:69:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 67 |             "bailing_moe": create(BailingMoeConfiguration.self, BailingMoeModel.init),
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:70:22: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 68 |             "lfm2_moe": create(LFM2MoEConfiguration.self, LFM2MoEModel.init),
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
    |                      `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:71:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 69 |             "nanochat": create(NanoChatConfiguration.self, NanoChatModel.init),
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
 73 |         ]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:72:25: warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 70 |             "afmoe": create(AfMoEConfiguration.self, AfMoEModel.init),
 71 |             "jamba_3b": create(JambaConfiguration.self, JambaModel.init),
 72 |             "mistral3": create(Mistral3TextConfiguration.self, Mistral3TextModel.init),
    |                         `- warning: converting non-Sendable function value to '@Sendable (URL) throws -> any LanguageModel' may introduce data races
 73 |         ]
 74 |     }
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/AfMoE.swift:502:22: warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
500 |         // Create attention masks
501 |         let faCache: [KVCache]? = layerCache.map { [$0[faIdx]] }
502 |         let faMask = createAttentionMask(h: h, cache: faCache)
    |                      `- warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
503 |
504 |         var swaMask: MLXFast.ScaledDotProductAttentionMaskMode = .none
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/LLMModelFactory.swift:525:17: warning: returning task-isolated 'model' as a 'sending' result risks causing data races; this is an error in the Swift 6 language mode
523 |             messageGenerator: messageGenerator)
524 |
525 |         return .init(
    |                 |- warning: returning task-isolated 'model' as a 'sending' result risks causing data races; this is an error in the Swift 6 language mode
    |                 `- note: returning task-isolated 'model' risks causing data races since the caller assumes that 'model' can be safely sent to other isolation domains
526 |             configuration: configuration, model: model, processor: processor, tokenizer: tokenizer)
527 |     }
[#DeprecatedDeclaration]: <https://docs.swift.org/compiler/documentation/diagnostics/deprecated-declaration>
[493/512] Compiling MLXLLM Gemma.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Gemma3Text.swift:13:8: warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 11 | import MLX
 12 | import MLXFast
 13 | import MLXLLM
    |        `- warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 14 | import MLXLMCommon
 15 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:124:13: warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
122 |
123 |     public init(_ args: GraniteConfiguration) {
124 |         let attentionHeads = args.attentionHeads
    |             `- warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
125 |         let hiddenSize = args.hiddenSize
126 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:274:74: warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
272 |         self.kvHeads = try container.decode(Int.self, forKey: .kvHeads)
273 |         self.attentionBias = try container.decode(Bool.self, forKey: .attentionBias)
274 |         self.mlpBias = try container.decode(Bool.self, forKey: .mlpBias) ?? false
    |                                                                          `- warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
275 |         self.ropeTheta = try container.decodeIfPresent(Float.self, forKey: .ropeTheta) ?? 10000000.0
276 |         self.ropeScaling = try container.decodeIfPresent(
[494/512] Compiling MLXLLM Gemma2.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Gemma3Text.swift:13:8: warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 11 | import MLX
 12 | import MLXFast
 13 | import MLXLLM
    |        `- warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 14 | import MLXLMCommon
 15 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:124:13: warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
122 |
123 |     public init(_ args: GraniteConfiguration) {
124 |         let attentionHeads = args.attentionHeads
    |             `- warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
125 |         let hiddenSize = args.hiddenSize
126 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:274:74: warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
272 |         self.kvHeads = try container.decode(Int.self, forKey: .kvHeads)
273 |         self.attentionBias = try container.decode(Bool.self, forKey: .attentionBias)
274 |         self.mlpBias = try container.decode(Bool.self, forKey: .mlpBias) ?? false
    |                                                                          `- warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
275 |         self.ropeTheta = try container.decodeIfPresent(Float.self, forKey: .ropeTheta) ?? 10000000.0
276 |         self.ropeScaling = try container.decodeIfPresent(
[495/512] Compiling MLXLLM Gemma3Text.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Gemma3Text.swift:13:8: warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 11 | import MLX
 12 | import MLXFast
 13 | import MLXLLM
    |        `- warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 14 | import MLXLMCommon
 15 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:124:13: warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
122 |
123 |     public init(_ args: GraniteConfiguration) {
124 |         let attentionHeads = args.attentionHeads
    |             `- warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
125 |         let hiddenSize = args.hiddenSize
126 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:274:74: warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
272 |         self.kvHeads = try container.decode(Int.self, forKey: .kvHeads)
273 |         self.attentionBias = try container.decode(Bool.self, forKey: .attentionBias)
274 |         self.mlpBias = try container.decode(Bool.self, forKey: .mlpBias) ?? false
    |                                                                          `- warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
275 |         self.ropeTheta = try container.decodeIfPresent(Float.self, forKey: .ropeTheta) ?? 10000000.0
276 |         self.ropeScaling = try container.decodeIfPresent(
[496/512] Compiling MLXLLM Gemma3nText.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Gemma3Text.swift:13:8: warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 11 | import MLX
 12 | import MLXFast
 13 | import MLXLLM
    |        `- warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 14 | import MLXLMCommon
 15 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:124:13: warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
122 |
123 |     public init(_ args: GraniteConfiguration) {
124 |         let attentionHeads = args.attentionHeads
    |             `- warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
125 |         let hiddenSize = args.hiddenSize
126 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:274:74: warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
272 |         self.kvHeads = try container.decode(Int.self, forKey: .kvHeads)
273 |         self.attentionBias = try container.decode(Bool.self, forKey: .attentionBias)
274 |         self.mlpBias = try container.decode(Bool.self, forKey: .mlpBias) ?? false
    |                                                                          `- warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
275 |         self.ropeTheta = try container.decodeIfPresent(Float.self, forKey: .ropeTheta) ?? 10000000.0
276 |         self.ropeScaling = try container.decodeIfPresent(
[497/512] Compiling MLXLLM Granite.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Gemma3Text.swift:13:8: warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 11 | import MLX
 12 | import MLXFast
 13 | import MLXLLM
    |        `- warning: file 'Gemma3Text.swift' is part of module 'MLXLLM'; ignoring import
 14 | import MLXLMCommon
 15 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:124:13: warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
122 |
123 |     public init(_ args: GraniteConfiguration) {
124 |         let attentionHeads = args.attentionHeads
    |             `- warning: initialization of immutable value 'attentionHeads' was never used; consider replacing with assignment to '_' or removing it [#no-usage]
125 |         let hiddenSize = args.hiddenSize
126 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Granite.swift:274:74: warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
272 |         self.kvHeads = try container.decode(Int.self, forKey: .kvHeads)
273 |         self.attentionBias = try container.decode(Bool.self, forKey: .attentionBias)
274 |         self.mlpBias = try container.decode(Bool.self, forKey: .mlpBias) ?? false
    |                                                                          `- warning: left side of nil coalescing operator '??' has non-optional type 'Bool', so the right side is never used
275 |         self.ropeTheta = try container.decodeIfPresent(Float.self, forKey: .ropeTheta) ?? 10000000.0
276 |         self.ropeScaling = try container.decodeIfPresent(
[498/512] Compiling MLXLLM Ernie4_5.swift
[499/512] Compiling MLXLLM Exaone4.swift
[500/512] Compiling MLXLLM FalconH1.swift
[501/512] Compiling MLXLLM GLM4.swift
[502/512] Compiling MLXLLM GPTOSS.swift
[503/512] Compiling MLXLLM BaichuanM1.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:6:8: warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  4 | import MLX
  5 | import MLXFast
  6 | import MLXLLM
    |        `- warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  7 | import MLXLMCommon
  8 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:231:17: warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
229 |         let inds = argPartition(-scores, kth: k - 1, axis: -1)[.ellipsis, ..<k]
230 |         scores = takeAlong(scores, inds, axis: -1)
231 |         if topK ?? 1 > 1, normTopkProb {
    |                 `- warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
232 |             let denominator = scores.sum(axis: -1, keepDims: true) + 1e-20
233 |             scores = scores / denominator
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:215:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
213 |
214 |     func groupSelect(_ x: MLXArray) -> (inds: MLXArray, scores: MLXArray) {
215 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
216 |
217 |         let logits = gate(x)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:224:13: warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
222 |         let topKGroup = top(groupScores, k: 2, axis: -1).sum(axis: -1, keepDims: true)
223 |         var k = nGroup - topkGroup
224 |         var groupIdx = argPartition(topKGroup, kth: k - 1, axis: -2)[.ellipsis, ..<k, 0...]
    |             `- warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
225 |         scores = putAlong(groupScores, groupIdx, values: MLXArray(0.0), axis: -2)
226 |         scores = flattened(scores, start: -2, end: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:357:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
355 |
356 |     func callAsFunction(_ x: MLXArray) -> (MLXArray, MLXArray) {
357 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
358 |
359 |         let hiddenStates = x.matmul(weight.T)
[504/512] Compiling MLXLLM BailingMoe.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:6:8: warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  4 | import MLX
  5 | import MLXFast
  6 | import MLXLLM
    |        `- warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  7 | import MLXLMCommon
  8 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:231:17: warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
229 |         let inds = argPartition(-scores, kth: k - 1, axis: -1)[.ellipsis, ..<k]
230 |         scores = takeAlong(scores, inds, axis: -1)
231 |         if topK ?? 1 > 1, normTopkProb {
    |                 `- warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
232 |             let denominator = scores.sum(axis: -1, keepDims: true) + 1e-20
233 |             scores = scores / denominator
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:215:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
213 |
214 |     func groupSelect(_ x: MLXArray) -> (inds: MLXArray, scores: MLXArray) {
215 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
216 |
217 |         let logits = gate(x)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:224:13: warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
222 |         let topKGroup = top(groupScores, k: 2, axis: -1).sum(axis: -1, keepDims: true)
223 |         var k = nGroup - topkGroup
224 |         var groupIdx = argPartition(topKGroup, kth: k - 1, axis: -2)[.ellipsis, ..<k, 0...]
    |             `- warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
225 |         scores = putAlong(groupScores, groupIdx, values: MLXArray(0.0), axis: -2)
226 |         scores = flattened(scores, start: -2, end: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:357:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
355 |
356 |     func callAsFunction(_ x: MLXArray) -> (MLXArray, MLXArray) {
357 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
358 |
359 |         let hiddenStates = x.matmul(weight.T)
[505/512] Compiling MLXLLM Bitnet.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:6:8: warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  4 | import MLX
  5 | import MLXFast
  6 | import MLXLLM
    |        `- warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  7 | import MLXLMCommon
  8 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:231:17: warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
229 |         let inds = argPartition(-scores, kth: k - 1, axis: -1)[.ellipsis, ..<k]
230 |         scores = takeAlong(scores, inds, axis: -1)
231 |         if topK ?? 1 > 1, normTopkProb {
    |                 `- warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
232 |             let denominator = scores.sum(axis: -1, keepDims: true) + 1e-20
233 |             scores = scores / denominator
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:215:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
213 |
214 |     func groupSelect(_ x: MLXArray) -> (inds: MLXArray, scores: MLXArray) {
215 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
216 |
217 |         let logits = gate(x)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:224:13: warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
222 |         let topKGroup = top(groupScores, k: 2, axis: -1).sum(axis: -1, keepDims: true)
223 |         var k = nGroup - topkGroup
224 |         var groupIdx = argPartition(topKGroup, kth: k - 1, axis: -2)[.ellipsis, ..<k, 0...]
    |             `- warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
225 |         scores = putAlong(groupScores, groupIdx, values: MLXArray(0.0), axis: -2)
226 |         scores = flattened(scores, start: -2, end: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:357:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
355 |
356 |     func callAsFunction(_ x: MLXArray) -> (MLXArray, MLXArray) {
357 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
358 |
359 |         let hiddenStates = x.matmul(weight.T)
[506/512] Compiling MLXLLM Cohere.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:6:8: warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  4 | import MLX
  5 | import MLXFast
  6 | import MLXLLM
    |        `- warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  7 | import MLXLMCommon
  8 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:231:17: warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
229 |         let inds = argPartition(-scores, kth: k - 1, axis: -1)[.ellipsis, ..<k]
230 |         scores = takeAlong(scores, inds, axis: -1)
231 |         if topK ?? 1 > 1, normTopkProb {
    |                 `- warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
232 |             let denominator = scores.sum(axis: -1, keepDims: true) + 1e-20
233 |             scores = scores / denominator
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:215:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
213 |
214 |     func groupSelect(_ x: MLXArray) -> (inds: MLXArray, scores: MLXArray) {
215 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
216 |
217 |         let logits = gate(x)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:224:13: warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
222 |         let topKGroup = top(groupScores, k: 2, axis: -1).sum(axis: -1, keepDims: true)
223 |         var k = nGroup - topkGroup
224 |         var groupIdx = argPartition(topKGroup, kth: k - 1, axis: -2)[.ellipsis, ..<k, 0...]
    |             `- warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
225 |         scores = putAlong(groupScores, groupIdx, values: MLXArray(0.0), axis: -2)
226 |         scores = flattened(scores, start: -2, end: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:357:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
355 |
356 |     func callAsFunction(_ x: MLXArray) -> (MLXArray, MLXArray) {
357 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
358 |
359 |         let hiddenStates = x.matmul(weight.T)
[507/512] Compiling MLXLLM DeepseekV3.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:6:8: warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  4 | import MLX
  5 | import MLXFast
  6 | import MLXLLM
    |        `- warning: file 'DeepseekV3.swift' is part of module 'MLXLLM'; ignoring import
  7 | import MLXLMCommon
  8 | import MLXNN
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:231:17: warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
229 |         let inds = argPartition(-scores, kth: k - 1, axis: -1)[.ellipsis, ..<k]
230 |         scores = takeAlong(scores, inds, axis: -1)
231 |         if topK ?? 1 > 1, normTopkProb {
    |                 `- warning: left side of nil coalescing operator '??' has non-optional type 'Int', so the right side is never used
232 |             let denominator = scores.sum(axis: -1, keepDims: true) + 1e-20
233 |             scores = scores / denominator
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:215:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
213 |
214 |     func groupSelect(_ x: MLXArray) -> (inds: MLXArray, scores: MLXArray) {
215 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
216 |
217 |         let logits = gate(x)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/BailingMoe.swift:224:13: warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
222 |         let topKGroup = top(groupScores, k: 2, axis: -1).sum(axis: -1, keepDims: true)
223 |         var k = nGroup - topkGroup
224 |         var groupIdx = argPartition(topKGroup, kth: k - 1, axis: -2)[.ellipsis, ..<k, 0...]
    |             `- warning: variable 'groupIdx' was never mutated; consider changing to 'let' constant
225 |         scores = putAlong(groupScores, groupIdx, values: MLXArray(0.0), axis: -2)
226 |         scores = flattened(scores, start: -2, end: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/DeepseekV3.swift:357:27: warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
355 |
356 |     func callAsFunction(_ x: MLXArray) -> (MLXArray, MLXArray) {
357 |         let (bsz, seqLen, h) = (x.dim(0), x.dim(1), x.dim(2))
    |                           `- warning: immutable value 'h' was never used; consider replacing with '_' or removing it [#no-usage]
358 |
359 |         let hiddenStates = x.matmul(weight.T)
[508/512] Compiling MLXLLM GraniteMoeHybrid.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:122:13: warning: variable 'padded' was never mutated; consider changing to 'let' constant
120 |         }
121 |
122 |         var padded = concatenated([convState!, input], axis: 1)
    |             `- warning: variable 'padded' was never mutated; consider changing to 'let' constant
123 |
124 |         if let cache {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:139:13: warning: variable 'projected' was never mutated; consider changing to 'let' constant
137 |         cache: MambaCache?
138 |     ) -> MLXArray {
139 |         var projected = inProj(hiddenStates)
    |             `- warning: variable 'projected' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:142:13: warning: variable 'gate' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
142 |         var gate = splits[0]
    |             `- warning: variable 'gate' was never mutated; consider changing to 'let' constant
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:144:13: warning: variable 'dt' was never mutated; consider changing to 'let' constant
142 |         var gate = splits[0]
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
    |             `- warning: variable 'dt' was never mutated; consider changing to 'let' constant
145 |
146 |         if let mask {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:554:25: warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
552 |                 let prefix = "model.layers.\(layerIndex).block_sparse_moe"
553 |                 guard
554 |                     var inputWeight = sanitized.removeValue(forKey: "\(prefix).input_linear.weight")
    |                         `- warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
555 |                 else { continue }
556 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:275:13: warning: variable 'newState' was never mutated; consider changing to 'let' constant
273 |         delta = softplus(dtProj(delta))
274 |
275 |         var newState = expandedDimensions(delta * x, axis: -1) * expandedDimensions(B, axis: -2)
    |             `- warning: variable 'newState' was never mutated; consider changing to 'let' constant
276 |         let dtA = exp(expandedDimensions(delta, axis: -1) * A)
277 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:454:24: warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
452 |         let cacheArray = cache ?? Array(repeating: nil, count: layers.count)
453 |
454 |         let attnMask = createAttentionMask(h: h, cache: [cacheArray[attnIdx]!])
    |                        `- warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
455 |
456 |         for (i, layer) in layers.enumerated() {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:209:13: warning: variable 'B' was never mutated; consider changing to 'let' constant
207 |         let BCx = inProj(x)
208 |         let parts = BCx.split(parts: 3, axis: -1)
209 |         var B = parts[0]
    |             `- warning: variable 'B' was never mutated; consider changing to 'let' constant
210 |         let C = parts[1]
211 |         let xComp = parts[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:296:13: warning: variable 'indices' was never mutated; consider changing to 'let' constant
294 |
295 |         let k = topK
296 |         var indices = argPartition(-gates, kth: k - 1, axis: -1)[.ellipsis, ..<k]
    |             `- warning: variable 'indices' was never mutated; consider changing to 'let' constant
297 |         var scores = takeAlong(gates, indices, axis: -1)
298 |         if normTopKProb {
[#DeprecatedDeclaration]: <https://docs.swift.org/compiler/documentation/diagnostics/deprecated-declaration>
[509/512] Compiling MLXLLM Internlm2.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:122:13: warning: variable 'padded' was never mutated; consider changing to 'let' constant
120 |         }
121 |
122 |         var padded = concatenated([convState!, input], axis: 1)
    |             `- warning: variable 'padded' was never mutated; consider changing to 'let' constant
123 |
124 |         if let cache {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:139:13: warning: variable 'projected' was never mutated; consider changing to 'let' constant
137 |         cache: MambaCache?
138 |     ) -> MLXArray {
139 |         var projected = inProj(hiddenStates)
    |             `- warning: variable 'projected' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:142:13: warning: variable 'gate' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
142 |         var gate = splits[0]
    |             `- warning: variable 'gate' was never mutated; consider changing to 'let' constant
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:144:13: warning: variable 'dt' was never mutated; consider changing to 'let' constant
142 |         var gate = splits[0]
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
    |             `- warning: variable 'dt' was never mutated; consider changing to 'let' constant
145 |
146 |         if let mask {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:554:25: warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
552 |                 let prefix = "model.layers.\(layerIndex).block_sparse_moe"
553 |                 guard
554 |                     var inputWeight = sanitized.removeValue(forKey: "\(prefix).input_linear.weight")
    |                         `- warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
555 |                 else { continue }
556 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:275:13: warning: variable 'newState' was never mutated; consider changing to 'let' constant
273 |         delta = softplus(dtProj(delta))
274 |
275 |         var newState = expandedDimensions(delta * x, axis: -1) * expandedDimensions(B, axis: -2)
    |             `- warning: variable 'newState' was never mutated; consider changing to 'let' constant
276 |         let dtA = exp(expandedDimensions(delta, axis: -1) * A)
277 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:454:24: warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
452 |         let cacheArray = cache ?? Array(repeating: nil, count: layers.count)
453 |
454 |         let attnMask = createAttentionMask(h: h, cache: [cacheArray[attnIdx]!])
    |                        `- warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
455 |
456 |         for (i, layer) in layers.enumerated() {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:209:13: warning: variable 'B' was never mutated; consider changing to 'let' constant
207 |         let BCx = inProj(x)
208 |         let parts = BCx.split(parts: 3, axis: -1)
209 |         var B = parts[0]
    |             `- warning: variable 'B' was never mutated; consider changing to 'let' constant
210 |         let C = parts[1]
211 |         let xComp = parts[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:296:13: warning: variable 'indices' was never mutated; consider changing to 'let' constant
294 |
295 |         let k = topK
296 |         var indices = argPartition(-gates, kth: k - 1, axis: -1)[.ellipsis, ..<k]
    |             `- warning: variable 'indices' was never mutated; consider changing to 'let' constant
297 |         var scores = takeAlong(gates, indices, axis: -1)
298 |         if normTopKProb {
[#DeprecatedDeclaration]: <https://docs.swift.org/compiler/documentation/diagnostics/deprecated-declaration>
[510/512] Compiling MLXLLM Jamba.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:122:13: warning: variable 'padded' was never mutated; consider changing to 'let' constant
120 |         }
121 |
122 |         var padded = concatenated([convState!, input], axis: 1)
    |             `- warning: variable 'padded' was never mutated; consider changing to 'let' constant
123 |
124 |         if let cache {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:139:13: warning: variable 'projected' was never mutated; consider changing to 'let' constant
137 |         cache: MambaCache?
138 |     ) -> MLXArray {
139 |         var projected = inProj(hiddenStates)
    |             `- warning: variable 'projected' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:142:13: warning: variable 'gate' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
142 |         var gate = splits[0]
    |             `- warning: variable 'gate' was never mutated; consider changing to 'let' constant
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:144:13: warning: variable 'dt' was never mutated; consider changing to 'let' constant
142 |         var gate = splits[0]
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
    |             `- warning: variable 'dt' was never mutated; consider changing to 'let' constant
145 |
146 |         if let mask {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:554:25: warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
552 |                 let prefix = "model.layers.\(layerIndex).block_sparse_moe"
553 |                 guard
554 |                     var inputWeight = sanitized.removeValue(forKey: "\(prefix).input_linear.weight")
    |                         `- warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
555 |                 else { continue }
556 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:275:13: warning: variable 'newState' was never mutated; consider changing to 'let' constant
273 |         delta = softplus(dtProj(delta))
274 |
275 |         var newState = expandedDimensions(delta * x, axis: -1) * expandedDimensions(B, axis: -2)
    |             `- warning: variable 'newState' was never mutated; consider changing to 'let' constant
276 |         let dtA = exp(expandedDimensions(delta, axis: -1) * A)
277 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:454:24: warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
452 |         let cacheArray = cache ?? Array(repeating: nil, count: layers.count)
453 |
454 |         let attnMask = createAttentionMask(h: h, cache: [cacheArray[attnIdx]!])
    |                        `- warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
455 |
456 |         for (i, layer) in layers.enumerated() {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:209:13: warning: variable 'B' was never mutated; consider changing to 'let' constant
207 |         let BCx = inProj(x)
208 |         let parts = BCx.split(parts: 3, axis: -1)
209 |         var B = parts[0]
    |             `- warning: variable 'B' was never mutated; consider changing to 'let' constant
210 |         let C = parts[1]
211 |         let xComp = parts[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:296:13: warning: variable 'indices' was never mutated; consider changing to 'let' constant
294 |
295 |         let k = topK
296 |         var indices = argPartition(-gates, kth: k - 1, axis: -1)[.ellipsis, ..<k]
    |             `- warning: variable 'indices' was never mutated; consider changing to 'let' constant
297 |         var scores = takeAlong(gates, indices, axis: -1)
298 |         if normTopKProb {
[#DeprecatedDeclaration]: <https://docs.swift.org/compiler/documentation/diagnostics/deprecated-declaration>
[511/512] Compiling MLXLLM LFM2.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:122:13: warning: variable 'padded' was never mutated; consider changing to 'let' constant
120 |         }
121 |
122 |         var padded = concatenated([convState!, input], axis: 1)
    |             `- warning: variable 'padded' was never mutated; consider changing to 'let' constant
123 |
124 |         if let cache {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:139:13: warning: variable 'projected' was never mutated; consider changing to 'let' constant
137 |         cache: MambaCache?
138 |     ) -> MLXArray {
139 |         var projected = inProj(hiddenStates)
    |             `- warning: variable 'projected' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:142:13: warning: variable 'gate' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
142 |         var gate = splits[0]
    |             `- warning: variable 'gate' was never mutated; consider changing to 'let' constant
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:144:13: warning: variable 'dt' was never mutated; consider changing to 'let' constant
142 |         var gate = splits[0]
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
    |             `- warning: variable 'dt' was never mutated; consider changing to 'let' constant
145 |
146 |         if let mask {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:554:25: warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
552 |                 let prefix = "model.layers.\(layerIndex).block_sparse_moe"
553 |                 guard
554 |                     var inputWeight = sanitized.removeValue(forKey: "\(prefix).input_linear.weight")
    |                         `- warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
555 |                 else { continue }
556 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:275:13: warning: variable 'newState' was never mutated; consider changing to 'let' constant
273 |         delta = softplus(dtProj(delta))
274 |
275 |         var newState = expandedDimensions(delta * x, axis: -1) * expandedDimensions(B, axis: -2)
    |             `- warning: variable 'newState' was never mutated; consider changing to 'let' constant
276 |         let dtA = exp(expandedDimensions(delta, axis: -1) * A)
277 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:454:24: warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
452 |         let cacheArray = cache ?? Array(repeating: nil, count: layers.count)
453 |
454 |         let attnMask = createAttentionMask(h: h, cache: [cacheArray[attnIdx]!])
    |                        `- warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
455 |
456 |         for (i, layer) in layers.enumerated() {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:209:13: warning: variable 'B' was never mutated; consider changing to 'let' constant
207 |         let BCx = inProj(x)
208 |         let parts = BCx.split(parts: 3, axis: -1)
209 |         var B = parts[0]
    |             `- warning: variable 'B' was never mutated; consider changing to 'let' constant
210 |         let C = parts[1]
211 |         let xComp = parts[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:296:13: warning: variable 'indices' was never mutated; consider changing to 'let' constant
294 |
295 |         let k = topK
296 |         var indices = argPartition(-gates, kth: k - 1, axis: -1)[.ellipsis, ..<k]
    |             `- warning: variable 'indices' was never mutated; consider changing to 'let' constant
297 |         var scores = takeAlong(gates, indices, axis: -1)
298 |         if normTopKProb {
[#DeprecatedDeclaration]: <https://docs.swift.org/compiler/documentation/diagnostics/deprecated-declaration>
[512/512] Compiling MLXLLM LFM2MoE.swift
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:122:13: warning: variable 'padded' was never mutated; consider changing to 'let' constant
120 |         }
121 |
122 |         var padded = concatenated([convState!, input], axis: 1)
    |             `- warning: variable 'padded' was never mutated; consider changing to 'let' constant
123 |
124 |         if let cache {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:139:13: warning: variable 'projected' was never mutated; consider changing to 'let' constant
137 |         cache: MambaCache?
138 |     ) -> MLXArray {
139 |         var projected = inProj(hiddenStates)
    |             `- warning: variable 'projected' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:142:13: warning: variable 'gate' was never mutated; consider changing to 'let' constant
140 |         let splits = split(
141 |             projected, indices: [intermediateSize, intermediateSize + convDim], axis: -1)
142 |         var gate = splits[0]
    |             `- warning: variable 'gate' was never mutated; consider changing to 'let' constant
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:144:13: warning: variable 'dt' was never mutated; consider changing to 'let' constant
142 |         var gate = splits[0]
143 |         var convInput = splits[1]
144 |         var dt = splits[2]
    |             `- warning: variable 'dt' was never mutated; consider changing to 'let' constant
145 |
146 |         if let mask {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/GraniteMoeHybrid.swift:554:25: warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
552 |                 let prefix = "model.layers.\(layerIndex).block_sparse_moe"
553 |                 guard
554 |                     var inputWeight = sanitized.removeValue(forKey: "\(prefix).input_linear.weight")
    |                         `- warning: variable 'inputWeight' was never mutated; consider changing to 'let' constant
555 |                 else { continue }
556 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:275:13: warning: variable 'newState' was never mutated; consider changing to 'let' constant
273 |         delta = softplus(dtProj(delta))
274 |
275 |         var newState = expandedDimensions(delta * x, axis: -1) * expandedDimensions(B, axis: -2)
    |             `- warning: variable 'newState' was never mutated; consider changing to 'let' constant
276 |         let dtA = exp(expandedDimensions(delta, axis: -1) * A)
277 |
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/Jamba.swift:454:24: warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
452 |         let cacheArray = cache ?? Array(repeating: nil, count: layers.count)
453 |
454 |         let attnMask = createAttentionMask(h: h, cache: [cacheArray[attnIdx]!])
    |                        `- warning: 'createAttentionMask(h:cache:returnArray:)' is deprecated: Use createAttentionMask(h:cache:windowSize:returnArray:) with a single cache instead [#DeprecatedDeclaration]
455 |
456 |         for (i, layer) in layers.enumerated() {
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:209:13: warning: variable 'B' was never mutated; consider changing to 'let' constant
207 |         let BCx = inProj(x)
208 |         let parts = BCx.split(parts: 3, axis: -1)
209 |         var B = parts[0]
    |             `- warning: variable 'B' was never mutated; consider changing to 'let' constant
210 |         let C = parts[1]
211 |         let xComp = parts[2]
/Users/admin/builder/spi-builder-workspace/Libraries/MLXLLM/Models/LFM2MoE.swift:296:13: warning: variable 'indices' was never mutated; consider changing to 'let' constant
294 |
295 |         let k = topK
296 |         var indices = argPartition(-gates, kth: k - 1, axis: -1)[.ellipsis, ..<k]
    |             `- warning: variable 'indices' was never mutated; consider changing to 'let' constant
297 |         var scores = takeAlong(gates, indices, axis: -1)
298 |         if normTopKProb {
[#DeprecatedDeclaration]: <https://docs.swift.org/compiler/documentation/diagnostics/deprecated-declaration>
Build complete! (49.40s)
Build complete.
{
  "dependencies" : [
    {
      "identity" : "mlx-swift",
      "requirement" : {
        "range" : [
          {
            "lower_bound" : "0.29.1",
            "upper_bound" : "0.30.0"
          }
        ]
      },
      "type" : "sourceControl",
      "url" : "https://github.com/ml-explore/mlx-swift"
    },
    {
      "identity" : "swift-transformers",
      "requirement" : {
        "range" : [
          {
            "lower_bound" : "1.1.0",
            "upper_bound" : "1.2.0"
          }
        ]
      },
      "type" : "sourceControl",
      "url" : "https://github.com/huggingface/swift-transformers"
    }
  ],
  "manifest_display_name" : "mlx-swift-lm",
  "name" : "mlx-swift-lm",
  "path" : "/Users/admin/builder/spi-builder-workspace",
  "platforms" : [
    {
      "name" : "macos",
      "version" : "14.0"
    },
    {
      "name" : "ios",
      "version" : "16.0"
    }
  ],
  "products" : [
    {
      "name" : "MLXLLM",
      "targets" : [
        "MLXLLM"
      ],
      "type" : {
        "library" : [
          "automatic"
        ]
      }
    },
    {
      "name" : "MLXVLM",
      "targets" : [
        "MLXVLM"
      ],
      "type" : {
        "library" : [
          "automatic"
        ]
      }
    },
    {
      "name" : "MLXLMCommon",
      "targets" : [
        "MLXLMCommon"
      ],
      "type" : {
        "library" : [
          "automatic"
        ]
      }
    },
    {
      "name" : "MLXEmbedders",
      "targets" : [
        "MLXEmbedders"
      ],
      "type" : {
        "library" : [
          "automatic"
        ]
      }
    }
  ],
  "targets" : [
    {
      "c99name" : "MLXVLM",
      "module_type" : "SwiftTarget",
      "name" : "MLXVLM",
      "path" : "Libraries/MLXVLM",
      "product_dependencies" : [
        "MLX",
        "MLXFast",
        "MLXNN",
        "MLXOptimizers",
        "MLXRandom",
        "Transformers"
      ],
      "product_memberships" : [
        "MLXVLM"
      ],
      "sources" : [
        "MediaProcessing.swift",
        "Models/FastVLM.swift",
        "Models/Gemma3.swift",
        "Models/Idefics3.swift",
        "Models/Paligemma.swift",
        "Models/Qwen25VL.swift",
        "Models/Qwen2VL.swift",
        "Models/Qwen3VL.swift",
        "Models/QwenVL.swift",
        "Models/SmolVLM2.swift",
        "VLMModel.swift",
        "VLMModelFactory.swift"
      ],
      "target_dependencies" : [
        "MLXLMCommon"
      ],
      "type" : "library"
    },
    {
      "c99name" : "MLXLMTests",
      "module_type" : "SwiftTarget",
      "name" : "MLXLMTests",
      "path" : "Tests/MLXLMTests",
      "product_dependencies" : [
        "MLX",
        "MLXNN",
        "MLXOptimizers",
        "MLXRandom",
        "Transformers"
      ],
      "sources" : [
        "BaseConfigurationTests.swift",
        "EvalTests.swift",
        "StreamlinedTests.swift",
        "ToolTests.swift",
        "UserInputTests.swift"
      ],
      "target_dependencies" : [
        "MLXLMCommon",
        "MLXLLM",
        "MLXVLM"
      ],
      "type" : "test"
    },
    {
      "c99name" : "MLXLMCommon",
      "module_type" : "SwiftTarget",
      "name" : "MLXLMCommon",
      "path" : "Libraries/MLXLMCommon",
      "product_dependencies" : [
        "MLX",
        "MLXNN",
        "MLXOptimizers",
        "MLXRandom",
        "MLXLinalg",
        "Transformers"
      ],
      "product_memberships" : [
        "MLXLLM",
        "MLXVLM",
        "MLXLMCommon",
        "MLXEmbedders"
      ],
      "sources" : [
        "Adapters/LoRA/DoRA+Layers.swift",
        "Adapters/LoRA/LoRA+Layers.swift",
        "Adapters/LoRA/LoRAContainer.swift",
        "Adapters/LoRA/LoRAModel.swift",
        "Adapters/ModelAdapter.swift",
        "Adapters/ModelAdapterFactory.swift",
        "Adapters/ModelAdapterTypeRegistry.swift",
        "AttentionUtils.swift",
        "BaseConfiguration.swift",
        "Chat.swift",
        "Evaluate.swift",
        "Extensions/Encodable+toolResult.swift",
        "KVCache.swift",
        "LanguageModel.swift",
        "Load.swift",
        "ModelConfiguration.swift",
        "ModelContainer.swift",
        "ModelFactory.swift",
        "Models/Gemma.swift",
        "Module+Extensions.swift",
        "Registries/AbstractModelRegistry.swift",
        "Registries/ModelTypeRegistry.swift",
        "Registries/ProcessorTypeRegistry.swift",
        "Streamlined.swift",
        "StringOrNumber.swift",
        "Tokenizer.swift",
        "Tool/Tool.swift",
        "Tool/ToolCall.swift",
        "Tool/ToolCallProcessor.swift",
        "Tool/ToolParameter.swift",
        "Tool/Value.swift",
        "UserInput.swift"
      ],
      "type" : "library"
    },
    {
      "c99name" : "MLXLLM",
      "module_type" : "SwiftTarget",
      "name" : "MLXLLM",
      "path" : "Libraries/MLXLLM",
      "product_dependencies" : [
        "MLX",
        "MLXFast",
        "MLXNN",
        "MLXOptimizers",
        "MLXRandom",
        "Transformers"
      ],
      "product_memberships" : [
        "MLXLLM"
      ],
      "sources" : [
        "LLMModel.swift",
        "LLMModelFactory.swift",
        "Lora+Data.swift",
        "LoraTrain.swift",
        "Models/AfMoE.swift",
        "Models/BaichuanM1.swift",
        "Models/BailingMoe.swift",
        "Models/Bitnet.swift",
        "Models/Cohere.swift",
        "Models/DeepseekV3.swift",
        "Models/Ernie4_5.swift",
        "Models/Exaone4.swift",
        "Models/FalconH1.swift",
        "Models/GLM4.swift",
        "Models/GPTOSS.swift",
        "Models/Gemma.swift",
        "Models/Gemma2.swift",
        "Models/Gemma3Text.swift",
        "Models/Gemma3nText.swift",
        "Models/Granite.swift",
        "Models/GraniteMoeHybrid.swift",
        "Models/Internlm2.swift",
        "Models/Jamba.swift",
        "Models/LFM2.swift",
        "Models/LFM2MoE.swift",
        "Models/Lille130m.swift",
        "Models/Llama.swift",
        "Models/MiMo.swift",
        "Models/Mistral3Text.swift",
        "Models/NanoChat.swift",
        "Models/Olmo2.swift",
        "Models/Olmo3.swift",
        "Models/OlmoE.swift",
        "Models/OpenELM.swift",
        "Models/Phi.swift",
        "Models/Phi3.swift",
        "Models/PhiMoE.swift",
        "Models/Qwen2.swift",
        "Models/Qwen3.swift",
        "Models/Qwen3MoE.swift",
        "Models/RoPEUtils.swift",
        "Models/SSM.swift",
        "Models/SmolLM3.swift",
        "Models/Starcoder2.swift",
        "SuScaledRoPE.swift",
        "SwitchLayers.swift"
      ],
      "target_dependencies" : [
        "MLXLMCommon"
      ],
      "type" : "library"
    },
    {
      "c99name" : "MLXEmbedders",
      "module_type" : "SwiftTarget",
      "name" : "MLXEmbedders",
      "path" : "Libraries/Embedders",
      "product_dependencies" : [
        "MLX",
        "MLXFast",
        "MLXNN",
        "Transformers",
        "MLXLinalg"
      ],
      "product_memberships" : [
        "MLXEmbedders"
      ],
      "sources" : [
        "BaseConfiguration.swift",
        "Bert.swift",
        "Configuration.swift",
        "EmbeddingModel.swift",
        "Load.swift",
        "Models.swift",
        "NomicBert.swift",
        "Pooling.swift",
        "Qwen3.swift",
        "Tokenizer.swift"
      ],
      "target_dependencies" : [
        "MLXLMCommon"
      ],
      "type" : "library"
    }
  ],
  "tools_version" : "5.9"
}
✅  Doc result (pending) reported
========================================
GenerateDocs
========================================
Generating docs at path:  $PWD/.docs/ml-explore/mlx-swift-lm/main
Repository:               ml-explore/mlx-swift-lm
Swift version used:       6.2
        .package(url: "https://github.com/apple/swift-docc-plugin", from: "1.3.0")
Target:                   MLXLLM
Extracting symbol information for 'MLXLLM'...
Finished extracting symbol information for 'MLXLLM'. (3.89s)
Building documentation for 'MLXLLM'...
warning: 'AceReason' doesn't exist at '/MLXLLM'
  --> Documentation.md:51:5-51:14
49 | - ``MiMoModel``
50 | - ``GLM4Model``
51 + - ``AceReason``
warning: 'ChatSession' doesn't exist at '/MLXLLM/evaluation'
  --> evaluation.md:15:53-15:64
13 |
14 | The second question actually refers to information (the location) from the first
15 + question -- this context is maintained inside the ``ChatSession`` object.
16 |
17 | If you need a one-shot prompt/response simply create a ``ChatSession``, evaluate
warning: 'ChatSession' doesn't exist at '/MLXLLM/evaluation'
  --> evaluation.md:17:58-17:69
15 | question -- this context is maintained inside the ``ChatSession`` object.
16 |
17 + If you need a one-shot prompt/response simply create a ``ChatSession``, evaluate
18 | the prompt and discard.  Multiple ``ChatSession`` instances could also be used
19 | (at the cost of the memory in the `KVCache`) to handle multiple streams of
warning: 'ChatSession' doesn't exist at '/MLXLLM/evaluation'
  --> evaluation.md:18:37-18:48
16 |
17 | If you need a one-shot prompt/response simply create a ``ChatSession``, evaluate
18 + the prompt and discard.  Multiple ``ChatSession`` instances could also be used
19 | (at the cost of the memory in the `KVCache`) to handle multiple streams of
20 | context.
warning: 'ChatSession' doesn't exist at '/MLXLLM/evaluation'
  --> evaluation.md:41:10-41:21
39 |
40 | This same API supports VLMs as well.  Simply present the image or video
41 + to the ``ChatSession``:
42 |
43 | ```swift
warning: 'ChatSession' doesn't exist at '/MLXLLM/evaluation'
  --> evaluation.md:62:7-62:18
60 | ## Advanced Usage
61 |
62 + The ``ChatSession`` has a number of parameters you can supply when creating it:
63 |
64 | - **instructions**: optional instructions to the chat session, e.g. describing what type of responses to give
warning: 'GenerateParameters' doesn't exist at '/MLXLLM/evaluation'
  --> evaluation.md:69:13-69:31
67 |     - or that the responses should be very brief
68 | - **generateParameters**: parameters that control the generation of output, e.g. token limits and temperature
69 +     - see ``GenerateParameters``
70 | - **processing**: optional media processing instructions
warning: 'fuse(model:layers:deQuantize:)' doesn't exist at '/MLXLLM/LoRATrain'
   --> ../LoraTrain.swift:103:9-103:39
101 | ///
102 | /// - ``saveLoRAWeights(model:url:)`` -- write the LoRA weights to a file
103 + /// - ``fuse(model:layers:deQuantize:)`` -- fuse the LoRA weights and convert back into the original model
104 | ///     architecture. These weights can be saved and reloaded with normal model handling code.
105 | /// - ``evaluate(model:dataset:loss:tokenizer:batchSize:batchCount:)``-- compute the test loss
warning: 'convert(model:layers:)' doesn't exist at '/MLXLLM/LoRATrain/saveLoRAWeights(model:url:)'
   --> ../LoraTrain.swift:211:13-211:35
209 |     ///
210 |     /// ### See Also
211 +     /// - ``convert(model:layers:)``
212 |     /// - ``loadLoRAWeights(model:url:)``
213 |     public static func saveLoRAWeights(model: Module, url: URL) throws {
warning: 'loadLoRAWeights(model:url:)' doesn't exist at '/MLXLLM/LoRATrain/saveLoRAWeights(model:url:)'
   --> ../LoraTrain.swift:212:13-212:40
210 |     /// ### See Also
211 |     /// - ``convert(model:layers:)``
212 +     /// - ``loadLoRAWeights(model:url:)``
213 |     public static func saveLoRAWeights(model: Module, url: URL) throws {
214 |         let parameters = Dictionary(
Finished building documentation for 'MLXLLM' (0.63s)
Generated documentation archive at:
  /Users/admin/builder/spi-builder-workspace/.docs/ml-explore/mlx-swift-lm/main
Fetching https://github.com/apple/swift-docc-plugin
Updating https://github.com/huggingface/swift-transformers
Updating https://github.com/apple/swift-numerics
Updating https://github.com/ml-explore/mlx-swift
Updated https://github.com/apple/swift-numerics (0.57s)
Updated https://github.com/ml-explore/mlx-swift (0.57s)
Updated https://github.com/huggingface/swift-transformers (0.57s)
[1/2158] Fetching swift-docc-plugin
Updating https://github.com/apple/swift-collections.git
Updating https://github.com/huggingface/swift-jinja.git
Fetched https://github.com/apple/swift-docc-plugin from cache (1.29s)
Computing version for https://github.com/huggingface/swift-transformers
Computed https://github.com/huggingface/swift-transformers at 1.1.4 (1.85s)
Updated https://github.com/apple/swift-collections.git (0.57s)
Updated https://github.com/huggingface/swift-jinja.git (0.57s)
Computing version for https://github.com/ml-explore/mlx-swift
Computed https://github.com/ml-explore/mlx-swift at 0.29.1 (0.59s)
Computing version for https://github.com/huggingface/swift-jinja.git
Computed https://github.com/huggingface/swift-jinja.git at 2.2.0 (0.51s)
Computing version for https://github.com/apple/swift-numerics
Computed https://github.com/apple/swift-numerics at 1.1.1 (0.52s)
Computing version for https://github.com/apple/swift-collections.git
Computed https://github.com/apple/swift-collections.git at 1.3.0 (0.64s)
Computing version for https://github.com/apple/swift-docc-plugin
Computed https://github.com/apple/swift-docc-plugin at 1.4.5 (2.15s)
Fetching https://github.com/swiftlang/swift-docc-symbolkit
[1/3593] Fetching swift-docc-symbolkit
Fetched https://github.com/swiftlang/swift-docc-symbolkit from cache (1.21s)
Computing version for https://github.com/swiftlang/swift-docc-symbolkit
Computed https://github.com/swiftlang/swift-docc-symbolkit at 1.0.0 (1.96s)
Creating working copy for https://github.com/apple/swift-docc-plugin
Working copy of https://github.com/apple/swift-docc-plugin resolved at 1.4.5
Creating working copy for https://github.com/swiftlang/swift-docc-symbolkit
Working copy of https://github.com/swiftlang/swift-docc-symbolkit resolved at 1.0.0
Building for debugging...
[0/8] Write sources
[0/8] Write snippet-extract-tool-entitlement.plist
[0/8] Write sources
[4/8] Write swift-version-49B95AFC49DCD68C.txt
[6/53] Compiling Snippets Snippet.swift
[7/53] Compiling Snippets SnippetParser.swift
[8/53] Emitting module SymbolKit
[9/57] Compiling SymbolKit DeclarationFragments.swift
[10/57] Compiling SymbolKit Fragment.swift
[11/57] Compiling SymbolKit FragmentKind.swift
[12/57] Compiling SymbolKit FunctionParameter.swift
[13/57] Compiling SymbolKit FunctionSignature.swift
[14/57] Compiling SymbolKit Symbol.swift
[15/57] Compiling SymbolKit SymbolKind.swift
[16/57] Compiling SymbolKit SymbolGraph.swift
[17/57] Compiling SymbolKit GraphCollector.swift
[18/57] Emitting module Snippets
[19/57] Compiling SymbolKit Mixin+Equals.swift
[20/57] Compiling SymbolKit Mixin+Hash.swift
[21/57] Compiling SymbolKit Mixin.swift
[22/57] Compiling SymbolKit LineList.swift
[23/57] Compiling SymbolKit Position.swift
[24/57] Compiling SymbolKit Relationship.swift
[25/57] Compiling SymbolKit RelationshipKind.swift
[26/57] Compiling SymbolKit SourceOrigin.swift
[27/57] Compiling SymbolKit GenericConstraints.swift
[28/57] Compiling SymbolKit Swift.swift
[29/57] Compiling SymbolKit Identifier.swift
[30/57] Compiling SymbolKit KindIdentifier.swift
[31/57] Compiling SymbolKit Location.swift
[32/57] Compiling SymbolKit Mutability.swift
[33/57] Compiling SymbolKit SourceRange.swift
[34/57] Compiling SymbolKit Metadata.swift
[35/57] Compiling SymbolKit Module.swift
[36/57] Compiling SymbolKit OperatingSystem.swift
[37/57] Compiling SymbolKit Platform.swift
[38/57] Compiling SymbolKit Names.swift
[39/57] Compiling SymbolKit SPI.swift
[40/57] Compiling SymbolKit Snippet.swift
[41/57] Compiling SymbolKit Extension.swift
[42/57] Compiling SymbolKit GenericConstraint.swift
[43/57] Compiling SymbolKit GenericParameter.swift
[44/57] Compiling SymbolKit Generics.swift
[45/57] Compiling SymbolKit Namespace.swift
[46/57] Compiling SymbolKit SemanticVersion.swift
[47/57] Compiling SymbolKit AccessControl.swift
[48/57] Compiling SymbolKit Availability.swift
[49/57] Compiling SymbolKit AvailabilityItem.swift
[50/57] Compiling SymbolKit Domain.swift
[51/57] Compiling SymbolKit UnifiedSymbol+Encodable.swift
[52/57] Compiling SymbolKit UnifiedSymbol.swift
[53/57] Compiling SymbolKit UnifiedSymbolGraph+Encodable.swift
[54/57] Compiling SymbolKit UnifiedSymbolGraph.swift
[55/61] Compiling snippet_extract SymbolGraph+Snippet.swift
[56/61] Compiling snippet_extract URL+Status.swift
[57/61] Compiling snippet_extract SnippetBuildCommand.swift
[58/61] Emitting module snippet_extract
[58/61] Write Objects.LinkFileList
[59/61] Linking snippet-extract-tool
[60/61] Applying snippet-extract-tool
Build of product 'snippet-extract' complete! (3.61s)
Building for debugging...
[0/1] Write swift-version-49B95AFC49DCD68C.txt
Build of target: 'MLXLLM' complete! (0.67s)
Target:                   MLXVLM
Extracting symbol information for 'MLXVLM'...
Finished extracting symbol information for 'MLXVLM'. (0.84s)
Building documentation for 'MLXVLM'...
warning: 'LLMModelFactory' doesn't exist at '/MLXVLM/VLMTypeRegistry'
  --> Libraries/MLXVLM/VLMModelFactory.swift:76:28-76:43
74 | /// Registry of model type, e.g 'llama', to functions that can instantiate the model from configuration.
75 | ///
76 + /// Typically called via ``LLMModelFactory/load(hub:configuration:progressHandler:)``.
77 | public class VLMTypeRegistry: ModelTypeRegistry, @unchecked Sendable {
78 |
Finished building documentation for 'MLXVLM' (0.40s)
Generated documentation archive at:
  /Users/admin/builder/spi-builder-workspace/.docs/ml-explore/mlx-swift-lm/main
Building for debugging...
[0/3] Write swift-version-49B95AFC49DCD68C.txt
Build of product 'snippet-extract' complete! (0.61s)
Building for debugging...
[0/1] Write swift-version-49B95AFC49DCD68C.txt
Build of target: 'MLXVLM' complete! (0.67s)
Successfully merged into /Users/admin/builder/spi-builder-workspace/.docs-staging/ml-explore/mlx-swift-lm/main/index/index.json
Successfully merged into /Users/admin/builder/spi-builder-workspace/.docs-staging/ml-explore/mlx-swift-lm/main/linkable-paths.json
Target:                   MLXLMCommon
Extracting symbol information for 'MLXLMCommon'...
Finished extracting symbol information for 'MLXLMCommon'. (0.92s)
Building documentation for 'MLXLMCommon'...
warning: 'LoRAConvertToLinear' doesn't exist at '/MLXLMCommon/LoRALinear'
  --> ../Adapters/LoRA/LoRA+Layers.swift:14:74-14:93
12 | ///
13 | /// - converting `Linear` or `QuantizedLinear` layers to ``LoRALinear`` / ``QLoRALinear``
14 + /// - converting ``LoRALinear`` back to `Linear` or `QuantizedLinear` (``LoRAConvertToLinear``)
15 | /// - implementing the LoRA evaluation
16 | ///
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/LoRALinear'
  --> ../Adapters/LoRA/LoRA+Layers.swift:19:46-19:55
17 | /// ``QLoRALinear`` is the equivalent class for `QuantizedLinear`.
18 | ///
19 + /// This is not typically used directly -- ``LoRATrain/convert(model:layers:)`` is used to
20 | /// add the adapter layers to a given model.
21 | ///
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/LoRALinear'
  --> ../Adapters/LoRA/LoRA+Layers.swift:26:9-26:18
24 | /// - [QLoRA: Efficient Finetuning of Quantized LLMs](https://arxiv.org/abs/2305.14314)
25 | /// - ``QLoRALinear``
26 + /// - ``LoRATrain/convert(model:layers:)``
27 | /// - ``LoRATrain/fuse(model:layers:deQuantize:)``
28 | public class LoRALinear: Linear, LoRALayer {
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/LoRALinear'
  --> ../Adapters/LoRA/LoRA+Layers.swift:27:9-27:18
25 | /// - ``QLoRALinear``
26 | /// - ``LoRATrain/convert(model:layers:)``
27 + /// - ``LoRATrain/fuse(model:layers:deQuantize:)``
28 | public class LoRALinear: Linear, LoRALayer {
29 |
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/LoRALinear/from(linear:rank:scale:)'
  --> ../Adapters/LoRA/LoRA+Layers.swift:69:40-69:49
67 |     /// that implements the `LoRA` adapter.
68 |     ///
69 +     /// This is typically called via ``LoRATrain/convert(model:layers:)``.
70 |     ///
71 |     /// ### See Also
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/LoRALinear/from(linear:rank:scale:)'
  --> ../Adapters/LoRA/LoRA+Layers.swift:72:13-72:22
70 |     ///
71 |     /// ### See Also
72 +     /// - ``LoRATrain/convert(model:layers:)``
73 |     /// - ``QLoRALinear/from(linear:rank:)``
74 |     public static func from(linear: Linear, rank: Int = 8, scale: Float = 20.0) -> LoRALayer {
warning: 'from(linear:rank:)' doesn't exist at '/MLXLMCommon/QLoRALinear'
  --> ../Adapters/LoRA/LoRA+Layers.swift:73:25-73:43
71 |     /// ### See Also
72 |     /// - ``LoRATrain/convert(model:layers:)``
73 +     /// - ``QLoRALinear/from(linear:rank:)``
   |                         ╰─suggestion: Replace 'from(linear:rank:)' with 'from(linear:rank:scale:)'
74 |     public static func from(linear: Linear, rank: Int = 8, scale: Float = 20.0) -> LoRALayer {
75 |         if let linear = linear as? QuantizedLinear {
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/LoRALinear/fused()'
  --> ../Adapters/LoRA/LoRA+Layers.swift:85:40-85:49
83 |     /// Convert back into a fused `Linear` layer.
84 |     ///
85 +     /// This is typically called via ``LoRATrain/fuse(model:layers:deQuantize:)``.
86 |     ///
87 |     /// ### See Also
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/LoRALinear/fused()'
  --> ../Adapters/LoRA/LoRA+Layers.swift:88:13-88:22
86 |     ///
87 |     /// ### See Also
88 +     /// - ``LoRATrain/fuse(model:layers:deQuantize:)``
89 |     /// - ``LoRAConvertToLinear``
90 |     /// - ``QLoRALinear/toLinear(deQuantize:)``
warning: 'LoRAConvertToLinear' doesn't exist at '/MLXLMCommon/LoRALinear/fused()'
  --> ../Adapters/LoRA/LoRA+Layers.swift:89:13-89:32
87 |     /// ### See Also
88 |     /// - ``LoRATrain/fuse(model:layers:deQuantize:)``
89 +     /// - ``LoRAConvertToLinear``
90 |     /// - ``QLoRALinear/toLinear(deQuantize:)``
91 |     public func fused() -> Module {
warning: 'toLinear(deQuantize:)' doesn't exist at '/MLXLMCommon/QLoRALinear'
  --> ../Adapters/LoRA/LoRA+Layers.swift:90:25-90:46
88 |     /// - ``LoRATrain/fuse(model:layers:deQuantize:)``
89 |     /// - ``LoRAConvertToLinear``
90 +     /// - ``QLoRALinear/toLinear(deQuantize:)``
91 |     public func fused() -> Module {
92 |         let dtype = weight.dtype
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/QLoRALinear/from(linear:rank:scale:)'
   --> ../Adapters/LoRA/LoRA+Layers.swift:153:40-153:49
151 |     /// that implements the `LoRA` adapter.
152 |     ///
153 +     /// This is typically called via ``LoRATrain/convert(model:layers:)``.
154 |     ///
155 |     /// ### See Also
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/QLoRALinear/from(linear:rank:scale:)'
   --> ../Adapters/LoRA/LoRA+Layers.swift:156:13-156:22
154 |     ///
155 |     /// ### See Also
156 +     /// - ``LoRATrain/convert(model:layers:)``
157 |     /// - ``LoRALinear/from(linear:rank:)``
158 |     public static func from(linear: QuantizedLinear, rank: Int = 8, scale: Float = 20.0)
warning: 'from(linear:rank:)' doesn't exist at '/MLXLMCommon/LoRALinear'
   --> ../Adapters/LoRA/LoRA+Layers.swift:157:24-157:42
155 |     /// ### See Also
156 |     /// - ``LoRATrain/convert(model:layers:)``
157 +     /// - ``LoRALinear/from(linear:rank:)``
    |                        ╰─suggestion: Replace 'from(linear:rank:)' with 'from(linear:rank:scale:)'
158 |     public static func from(linear: QuantizedLinear, rank: Int = 8, scale: Float = 20.0)
159 |         -> LoRALayer
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/QLoRALinear/fused()'
   --> ../Adapters/LoRA/LoRA+Layers.swift:168:40-168:49
166 |     /// Convert back into a fused `QuantizedLinear` layer.
167 |     ///
168 +     /// This is typically called via ``LoRATrain/fuse(model:layers:deQuantize:)``.
169 |     ///
170 |     /// ### See Also
warning: 'LoRATrain' doesn't exist at '/MLXLMCommon/QLoRALinear/fused()'
   --> ../Adapters/LoRA/LoRA+Layers.swift:171:13-171:22
169 |     ///
170 |     /// ### See Also
171 +     /// - ``LoRATrain/fuse(model:layers:deQuantize:)``
172 |     public func fused() -> Module {
173 |         let weight = dequantizedWeight
warning: 'Chat.Message' doesn't exist at '/MLXLMCommon/MessageGenerator'
  --> ../Chat.swift:59:7-59:19
57 |
58 | /// Protocol for something that can convert structured
59 + /// ``Chat.Message`` into model specific ``Message``
60 | /// (raw dictionary) format.
61 | ///
warning: 'loadWeights(modelDirectory:model:quantization:)' doesn't exist at '/MLXLMCommon/porting'
   --> porting.md:298:189-298:236
296 | ```
297 |
298 + Note that the weight is given an initial value and shape based on the parameters to the initializer. In typical inference use, these values will be replaced when the weights are loaded (``loadWeights(modelDirectory:model:quantization:)``).
299 |
300 | * Note: If the property names in Python don't make good Swift names, you can use the `@ParameterInfo` property wrapper to specify the key: `@ParameterInfo(key: "some_weight") var weight: MLXArray`
warning: 'generate(input:parameters:context:didGenerate:)' is ambiguous at '/MLXLMCommon'
   --> ../Evaluate.swift:249:44-249:91
247 | /// Generator of tokens.
248 | ///
249 + /// This is typically used via a call to ``generate(input:parameters:context:didGenerate:)``.
    |                                                                                           ├─suggestion: Insert '->GenerateCompletionInfo' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: (Int) -> GenerateDisposition) throws -> GenerateCompletionInfo'
    |                                                                                           ╰─suggestion: Insert '->GenerateResult' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: ([Int]) -> GenerateDisposition) throws -> GenerateResult'
250 | ///
251 | /// To use it directly:
warning: 'init(input:model:cache:processor:sampler:prefillStepSize:)' doesn't exist at '/MLXLMCommon/TokenIterator/init(input:model:cache:parameters:)'
   --> ../Evaluate.swift:323:11-323:69
321 |     ///
322 |     /// If more control is needed over the generation,
323 +     /// ``init(input:model:cache:processor:sampler:prefillStepSize:)``
324 |     /// allows a caller to specify ``LogitProcessor`` and ``LogitSampler``
325 |     /// directly.
warning: 'generate(input:parameters:context:didGenerate:)' is ambiguous at '/MLXLMCommon'
   --> ../Evaluate.swift:455:27-455:74
453 | }
454 |
455 + /// Result of a call to ``generate(input:parameters:context:didGenerate:)``.
    |                                                                          ├─suggestion: Insert '->GenerateCompletionInfo' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: (Int) -> GenerateDisposition) throws -> GenerateCompletionInfo'
    |                                                                          ╰─suggestion: Insert '->GenerateResult' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: ([Int]) -> GenerateDisposition) throws -> GenerateResult'
456 | public struct GenerateResult: Sendable {
457 |
warning: 'generate(input:parameters:context:didGenerate:)' is ambiguous at '/MLXLMCommon'
   --> ../Evaluate.swift:521:45-521:92
519 | }
520 |
521 + /// Action from token visitor callback in ``generate(input:parameters:context:didGenerate:)``.
    |                                                                                            ├─suggestion: Insert '->GenerateCompletionInfo' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: (Int) -> GenerateDisposition) throws -> GenerateCompletionInfo'
    |                                                                                            ╰─suggestion: Insert '->GenerateResult' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: ([Int]) -> GenerateDisposition) throws -> GenerateResult'
522 | public enum GenerateDisposition: Sendable {
523 |     /// keep producing tokens until an EOS token is produced
warning: 'generate(input:parameters:context:didGenerate:)' is ambiguous at '/MLXLMCommon'
   --> ../Evaluate.swift:532:7-532:54
530 | /// Given prompt tokens generate text using the given model and parameters.
531 | ///
532 + /// ``generate(input:parameters:context:didGenerate:)`` is the preferred call.
    |                                                      ├─suggestion: Insert '->GenerateCompletionInfo' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: (Int) -> GenerateDisposition) throws -> GenerateCompletionInfo'
    |                                                      ╰─suggestion: Insert '->GenerateResult' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: ([Int]) -> GenerateDisposition) throws -> GenerateResult'
533 | ///
534 | /// - Parameters:
warning: 'generate(input:context:iterator:didGenerate:)' is ambiguous at '/MLXLMCommon'
   --> ../Evaluate.swift:582:7-582:52
580 | ///
581 | /// Internally this constructs a ``TokenIterator`` and calls
582 + /// ``generate(input:context:iterator:didGenerate:)``
    |                                                    ├─suggestion: Insert '->GenerateCompletionInfo' for 'func generate(input: LMInput, context: ModelContext, iterator: TokenIterator, didGenerate: (Int) -> GenerateDisposition) -> GenerateCompletionInfo'
    |                                                    ╰─suggestion: Insert '->GenerateResult' for 'func generate(input: LMInput, context: ModelContext, iterator: TokenIterator, didGenerate: ([Int]) -> GenerateDisposition) -> GenerateResult'
583 | ///
584 | /// - Parameters:
warning: 'generate(input:parameters:context:didGenerate:)' is ambiguous at '/MLXLMCommon'
   --> ../Evaluate.swift:602:7-602:54
600 | /// Low level token generation using a ``TokenIterator``.
601 | ///
602 + /// ``generate(input:parameters:context:didGenerate:)`` is the preferred call.
    |                                                      ├─suggestion: Insert '->GenerateCompletionInfo' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: (Int) -> GenerateDisposition) throws -> GenerateCompletionInfo'
    |                                                      ╰─suggestion: Insert '->GenerateResult' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: ([Int]) -> GenerateDisposition) throws -> GenerateResult'
603 | ///
604 | /// - Parameters:
warning: 'generate(input:context:iterator:didGenerate:)' is ambiguous at '/MLXLMCommon'
   --> ../Evaluate.swift:681:7-681:52
679 | ///
680 | /// Internally this constructs a ``TokenIterator`` and calls
681 + /// ``generate(input:context:iterator:didGenerate:)``
    |                                                    ├─suggestion: Insert '->GenerateCompletionInfo' for 'func generate(input: LMInput, context: ModelContext, iterator: TokenIterator, didGenerate: (Int) -> GenerateDisposition) -> GenerateCompletionInfo'
    |                                                    ╰─suggestion: Insert '->GenerateResult' for 'func generate(input: LMInput, context: ModelContext, iterator: TokenIterator, didGenerate: ([Int]) -> GenerateDisposition) -> GenerateResult'
682 | ///
683 | /// - Parameters:
warning: 'MultiHeadAttention' doesn't exist at '/MLXLMCommon/createAttentionMask(h:cache:)'
   --> ../KVCache.swift:200:16-200:34
198 | /// Create an attention mask using the parameters from the KVCache.
199 | ///
200 + /// See also ``MultiHeadAttention/createAdditiveCausalMask(_:dtype:)`` -- same idea
201 | /// but doesn't honor the cache offset.
202 | @_disfavoredOverload
warning: 'generate(input:parameters:context:didGenerate:)' is ambiguous at '/MLXLMCommon'
  --> ../ModelFactory.swift:57:52-57:99
55 | ///
56 | /// - ``ModelConfiguration`` -- identifier for the model
57 + /// - ``LanguageModel`` -- the model itself, see ``generate(input:parameters:context:didGenerate:)``
   |                                                                                                   ├─suggestion: Insert '->GenerateCompletionInfo' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: (Int) -> GenerateDisposition) throws -> GenerateCompletionInfo'
   |                                                                                                   ╰─suggestion: Insert '->GenerateResult' for 'func generate(input: LMInput, parameters: GenerateParameters, context: ModelContext, didGenerate: ([Int]) -> GenerateDisposition) throws -> GenerateResult'
58 | /// - ``UserInputProcessor`` -- can convert ``UserInput`` into ``LMInput``
59 | /// - `Tokenizer` -- the tokenizer used by ``UserInputProcessor``
warning: 'loadModel(hub:id:progressHandler:)' doesn't exist at '/MLXLMCommon/ModelFactory'
  --> ../ModelFactory.swift:83:9-83:43
81 | ///
82 | /// ## See Also
83 + /// - ``loadModel(hub:id:progressHandler:)``
   |         ├─suggestion: Replace 'loadModel(hub:id:progressHandler:)' with 'load(hub:configuration:progressHandler:)'
   |         ╰─suggestion: Replace 'loadModel(hub:id:progressHandler:)' with 'loadContainer(hub:configuration:progressHandler:)'
84 | /// - ``loadModel(hub:directory:progressHandler:)``
85 | /// - ``loadModelContainer(hub:id:progressHandler:)``
warning: 'loadModelContainer(hub:id:progressHandler:)' doesn't exist at '/MLXLMCommon/ModelFactory'
  --> ../ModelFactory.swift:85:9-85:52
83 | /// - ``loadModel(hub:id:progressHandler:)``
84 | /// - ``loadModel(hub:directory:progressHandler:)``
85 + /// - ``loadModelContainer(hub:id:progressHandler:)``
   |         ├─suggestion: Replace 'loadModelContainer(hub:id:progressHandler:)' with 'loadContainer(hub:configuration:progressHandler:)'
   |         ╰─suggestion: Replace 'loadModelContainer(hub:id:progressHandler:)' with 'load(hub:configuration:progressHandler:)'
86 | /// - ``loadModelContainer(hub:directory:progressHandler:)``
87 | public protocol ModelFactory: Sendable {
warning: 'loadModel(hub:id:progressHandler:)' doesn't exist at '/MLXLMCommon/ModelFactory/load(hub:configuration:progressHandler:)'
   --> ../ModelFactory.swift:138:13-138:47
136 |     ///
137 |     /// ## See Also
138 +     /// - ``loadModel(hub:id:progressHandler:)``
139 |     /// - ``loadModelContainer(hub:id:progressHandler:)``
140 |     public func load(
warning: 'loadModelContainer(hub:id:progressHandler:)' doesn't exist at '/MLXLMCommon/ModelFactory/load(hub:configuration:progressHandler:)'
   --> ../ModelFactory.swift:139:13-139:56
137 |     /// ## See Also
138 |     /// - ``loadModel(hub:id:progressHandler:)``
139 +     /// - ``loadModelContainer(hub:id:progressHandler:)``
140 |     public func load(
141 |         hub: HubApi = defaultHubApi, configuration: ModelConfiguration,
warning: Parameter 'revision' is missing documentation
   --> ../ModelFactory.swift:216:56-216:56
214 | ///   - hub: optional HubApi -- by default uses ``defaultHubApi``
215 | ///   - id: huggingface model identifier, e.g "mlx-community/Qwen3-4B-4bit"
216 + ///   - progressHandler: optional callback for progress
    |       ╰─suggestion: Document 'revision' parameter
217 | /// - Returns: a ``ModelContext``
218 | public func loadModel(
warning: Parameter 'revision' is missing documentation
   --> ../ModelFactory.swift:237:56-237:56
235 | ///   - hub: optional HubApi -- by default uses ``defaultHubApi``
236 | ///   - id: huggingface model identifier, e.g "mlx-community/Qwen3-4B-4bit"
237 + ///   - progressHandler: optional callback for progress
    |       ╰─suggestion: Document 'revision' parameter
238 | /// - Returns: a ``ModelContainer``
239 | public func loadModelContainer(
warning: 'loadModel(hub:id:progressHandler:)' doesn't exist at '/MLXLMCommon/ModelFactoryRegistry'
   --> ../ModelFactory.swift:342:19-342:53
340 | /// Registry of ``ModelFactory`` trampolines.
341 | ///
342 + /// This allows ``loadModel(hub:id:progressHandler:)`` to use any ``ModelFactory`` instances
343 | /// available but be defined in the `LLMCommon` layer.  This is not typically used directly -- it is
344 | /// called via ``loadModel(hub:id:progressHandler:)``:
warning: 'loadModel(hub:id:progressHandler:)' doesn't exist at '/MLXLMCommon/ModelFactoryRegistry'
   --> ../ModelFactory.swift:344:18-344:52
342 | /// This allows ``loadModel(hub:id:progressHandler:)`` to use any ``ModelFactory`` instances
343 | /// available but be defined in the `LLMCommon` layer.  This is not typically used directly -- it is
344 + /// called via ``loadModel(hub:id:progressHandler:)``:
345 | ///
346 | /// ```swift
warning: 'loadModel(hub:id:progressHandler:)' doesn't exist at '/MLXLMCommon/ModelFactoryRegistry'
   --> ../ModelFactory.swift:351:9-351:43
349 | ///
350 | /// ## See Also
351 + /// - ``loadModel(hub:id:progressHandler:)``
352 | /// - ``loadModel(hub:directory:progressHandler:)``
353 | /// - ``loadModelContainer(hub:id:progressHandler:)``
warning: 'loadModelContainer(hub:id:progressHandler:)' doesn't exist at '/MLXLMCommon/ModelFactoryRegistry'
   --> ../ModelFactory.swift:353:9-353:52
351 | /// - ``loadModel(hub:id:progressHandler:)``
352 | /// - ``loadModel(hub:directory:progressHandler:)``
353 + /// - ``loadModelContainer(hub:id:progressHandler:)``
354 | /// - ``loadModelContainer(hub:directory:progressHandler:)``
355 | final public class ModelFactoryRegistry: @unchecked Sendable {
warning: 'modelRegistry' doesn't exist at '/MLXLMCommon/AbstractModelRegistry/configuration(id:)'
  --> ../Registries/AbstractModelRegistry.swift:28:38-28:51
26 |     }
27 |
28 +     /// Returns configuration from ``modelRegistry``.
29 |     ///
30 |     /// - Note: If the id doesn't exists in the configuration, this will return a new instance of it.
warning: Parameter 'images' not found in instance method declaration
   --> ../Streamlined.swift:221:13-221:19
219 |     /// - Parameters:
220 |     ///   - prompt: the prompt
221 +     ///   - images: optional image (for use with VLMs)
    |             ╰─suggestion: Replace 'images' with 'image'
222 |     ///   - videos: optional video (for use with VLMs)
223 |     /// - Returns: response from the model
warning: Parameter 'videos' not found in instance method declaration
   --> ../Streamlined.swift:222:13-222:19
220 |     ///   - prompt: the prompt
221 |     ///   - images: optional image (for use with VLMs)
222 +     ///   - videos: optional video (for use with VLMs)
    |             ╰─suggestion: Replace 'videos' with 'video'
223 |     /// - Returns: response from the model
224 |     public func respond(
warning: Parameter 'video' is missing documentation
   --> ../Streamlined.swift:222:55-222:55
220 |     ///   - prompt: the prompt
221 |     ///   - images: optional image (for use with VLMs)
222 +     ///   - videos: optional video (for use with VLMs)
    |                                                       ╰─suggestion: Document 'video' parameter
223 |     /// - Returns: response from the model
224 |     public func respond(
warning: Parameter 'image' is missing documentation
   --> ../Streamlined.swift:222:55-222:55
220 |     ///   - prompt: the prompt
221 |     ///   - images: optional image (for use with VLMs)
222 +     ///   - videos: optional video (for use with VLMs)
    |                                                       ╰─suggestion: Document 'image' parameter
223 |     /// - Returns: response from the model
224 |     public func respond(
warning: 'init(chat:tools:additionalContext:)' doesn't exist at '/MLXLMCommon/UserInput/init(prompt:images:videos:tools:additionalContext:)'
   --> ../UserInput.swift:178:13-178:48
176 |     /// ### See Also
177 |     /// - ``Prompt-swift.enum/text(_:)``
178 +     /// - ``init(chat:tools:additionalContext:)``
179 |     public init(
180 |         prompt: String, images: [Image] = [Image](), videos: [Video] = [Video](),
warning: 'init(chat:tools:additionalContext:)' doesn't exist at '/MLXLMCommon/UserInput/init(messages:images:videos:tools:additionalContext:)'
   --> ../UserInput.swift:212:25-212:60
210 |     /// ```
211 |     ///
212 +     /// Typically the ``init(chat:tools:additionalContext:)`` should be used instead
213 |     /// along with a model specific ``MessageGenerator`` (supplied by the ``UserInputProcessor``).
214 |     ///
warning: 'init(chat:tools:additionalContext:)' doesn't exist at '/MLXLMCommon/UserInput/init(messages:images:videos:tools:additionalContext:)'
   --> ../UserInput.swift:223:13-223:48
221 |     /// ### See Also
222 |     /// - ``Prompt-swift.enum/text(_:)``
223 +     /// - ``init(chat:tools:additionalContext:)``
224 |     public init(
225 |         messages: [Message], images: [Image] = [Image](), videos: [Video] = [Video](),
warning: 'init(chat:tools:additionalContext:)' doesn't exist at '/MLXLMCommon/UserInput/init(chat:processing:tools:additionalContext:)'
   --> ../UserInput.swift:258:13-258:48
256 |     /// ### See Also
257 |     /// - ``Prompt-swift.enum/text(_:)``
258 +     /// - ``init(chat:tools:additionalContext:)``
259 |     public init(
260 |         chat: [Chat.Message],
warning: 'init(chat:tools:additionalContext:)' doesn't exist at '/MLXLMCommon/UserInput/init(prompt:images:videos:processing:tools:additionalContext:)'
   --> ../UserInput.swift:282:11-282:46
280 |     /// Initialize the `UserInput` with a preconfigured ``Prompt-swift.enum``.
281 |     ///
282 +     /// ``init(chat:tools:additionalContext:)`` is the preferred mechanism.
283 |     ///
284 |     /// - Parameters:
warning: 'init(chat:tools:additionalContext:)' doesn't exist at '/MLXLMCommon/UserInput/init(prompt:images:videos:processing:tools:additionalContext:)'
   --> ../UserInput.swift:293:13-293:48
291 |     /// ### See Also
292 |     /// - ``Prompt-swift.enum/text(_:)``
293 +     /// - ``init(chat:tools:additionalContext:)``
294 |     public init(
295 |         prompt: Prompt,
Finished building documentation for 'MLXLMCommon' (0.43s)
Generated documentation archive at:
  /Users/admin/builder/spi-builder-workspace/.docs/ml-explore/mlx-swift-lm/main
Building for debugging...
[0/3] Write swift-version-49B95AFC49DCD68C.txt
Build of product 'snippet-extract' complete! (0.68s)
Building for debugging...
[0/1] Write swift-version-49B95AFC49DCD68C.txt
Build of target: 'MLXLMCommon' complete! (0.75s)
Successfully merged into /Users/admin/builder/spi-builder-workspace/.docs-staging/ml-explore/mlx-swift-lm/main/index/index.json
Successfully merged into /Users/admin/builder/spi-builder-workspace/.docs-staging/ml-explore/mlx-swift-lm/main/linkable-paths.json
Target:                   MLXEmbedders
Extracting symbol information for 'MLXEmbedders'...
Finished extracting symbol information for 'MLXEmbedders'. (0.87s)
Building documentation for 'MLXEmbedders'...
warning: 'LanguageModel' doesn't exist at '/MLXEmbedders/BaseConfiguration'
 --> Libraries/Embedders/BaseConfiguration.swift:6:12-6:25
4 | import MLX
5 |
6 + /// Base ``LanguageModel`` configuration -- provides `modelType`
7 | /// and `quantization` (used in loading the model).
8 | ///
warning: 'ModelFactory' doesn't exist at '/MLXEmbedders/BaseConfiguration'
  --> Libraries/Embedders/BaseConfiguration.swift:9:23-9:35
7  | /// and `quantization` (used in loading the model).
8  | ///
9  + /// This is used by ``ModelFactory/load(hub:configuration:progressHandler:)``
10 | /// to determine the type of model to load.
11 | public struct BaseConfiguration: Codable, Sendable {
Finished building documentation for 'MLXEmbedders' (0.13s)
Generated documentation archive at:
  /Users/admin/builder/spi-builder-workspace/.docs/ml-explore/mlx-swift-lm/main
Building for debugging...
[0/3] Write swift-version-49B95AFC49DCD68C.txt
Build of product 'snippet-extract' complete! (0.66s)
Building for debugging...
[0/1] Write swift-version-49B95AFC49DCD68C.txt
Build of target: 'MLXEmbedders' complete! (0.73s)
Successfully merged into /Users/admin/builder/spi-builder-workspace/.docs-staging/ml-explore/mlx-swift-lm/main/index/index.json
Successfully merged into /Users/admin/builder/spi-builder-workspace/.docs-staging/ml-explore/mlx-swift-lm/main/linkable-paths.json
    5508
29	/Users/admin/builder/spi-builder-workspace/.docs/ml-explore/mlx-swift-lm/main
✅  Doc result (uploading) reported
========================================
SyncDocs
========================================
Doc archive source directory: $PWD/.docs/ml-explore/mlx-swift-lm/main
File count: 5508
Doc size:   29.0MB
Preparing doc bundle ...
Uploading prod-ml-explore-mlx-swift-lm-main-b7074dbf.zip to s3://spi-docs-inbox/prod-ml-explore-mlx-swift-lm-main-b7074dbf.zip
Copying... [10%]
Copying... [20%]
Copying... [30%]
Copying... [40%]
Copying... [50%]
Copying... [60%]
Copying... [70%]
Copying... [80%]
Copying... [90%]
Copying... [100%]
Done.