feat(iOS): 更新MNN后端模型配置优化性能

将MNN主模型从Qwen3.5-4B(~2.64GiB)降级为Qwen3.5-2B(~1.1GiB),因为4B版本
实测运行过慢,影响用户体验。iPhone17+/SME2设备使用2B模型,保留MLX
兜底方案用于模拟器和备用场景,确保AI推理性能和存储效率的平衡。
```
This commit is contained in:
link2026
2026-06-09 22:20:07 +08:00
parent ca5a3fa38b
commit b79ae54b7b
40 changed files with 1327 additions and 452 deletions

View File

@@ -3,7 +3,7 @@ import Foundation
/// MNN(CPU / SME2), `MNNLLMBridge`
/// `LLMSession`/`VLSession` actor ; `AIRuntime`
///
/// () Qwen3.5-4B MNN :`generate` ,
/// () Qwen3.5-2B MNN :`generate` ,
/// `analyze` <img> Omni imread ( OMNI ,xcframework )
/// ,; MNN,VL 退 MLX( `AIRuntime`)
actor MNNBackend {