Crate candle_coreml 

Re-exports

pub use builder::CoreMLModelBuilder;

pub use cache::CacheManager;

pub use config::ComponentConfig;

pub use config::Config;

pub use config::ConfigGenerator;

pub use config::ModelConfig;

pub use config::NamingConfig;

pub use config::ShapeConfig;

pub use config::TensorConfig;

pub use model::CoreMLModel;

pub use qwen::ModelNamingConfig;

pub use qwen::QwenConfig;

pub use qwen::QwenModel;

pub use state::CoreMLState;

pub use unified_model_loader::CachedModelInfo;

pub use unified_model_loader::UnifiedModelLoader;

pub use download::download_model;

pub use download::download_model_to;

pub use download::ensure_model_downloaded;

pub use download::get_cached_model_path;

pub use download::download_hf_model_clean;

pub use download::verify_download_completeness;

pub use download::CleanDownloadConfig;

pub use utils::mask;

pub use utils::multi_component;

pub use utils::sampling;

Modules

builder

CoreML model builder for convenient model loading

cache

Cache module for candle-coreml

cache_manager

clean_git_lfs_downloader

config

Configuration module for candle-coreml

config_generator

conversion

Tensor conversion utilities for CoreML integration

download

Download module for candle-coreml

model

Core CoreML model implementation

model_config

model_downloader

pipeline

Generic CoreML pipeline abstraction (model-agnostic)

qwen

Qwen model integration for candle-coreml

state

CoreML state management for autoregressive inference

unified_model_loader

Unified model loader that combines downloading, config generation, and model loading

utils

Shared utilities for transformer models and multi-component architectures

Functions

get_local_or_remote_file

Helper function to get a file locally first, then download from HuggingFace Hub if needed. Follows the same pattern as quantized-t5 example.