Crate llama_core

Llama Core, abbreviated as llama-core, defines a set of APIs. Developers can utilize these APIs to build applications based on large models, such as chatbots, RAG, and more.

Re-exports

• pub use error::LlamaCoreError;
• pub use graph::EngineType;
• pub use graph::Graph;
• pub use graph::GraphBuilder;

Modules

• audio
  Define APIs for audio generation, transcription, and translation.
• chat
  Define APIs for chat completion.
• completions
  Define APIs for completions.
• embeddings
  Define APIs for computing embeddings.
• error
  Error types for the Llama Core library.
• graph
  Define Graph and GraphBuilder APIs for creating a new computation graph.
• images
  Define APIs for image generation and edit.
• models
  Define APIs for querying models.
• rag
  Define APIs for RAG operations.
• utils
  Define utility functions.

Structs

• Metadata
  Model metadata
• MetadataBuilder
  Builder for the Metadata struct
• PluginInfo
  Version info of the wasi-nn_ggml plugin, including the build number and the commit id.
• WhisperMetadataBuilder
  Builder for creating an audio metadata

Enums

• RunningMode
  Running mode

Functions

• get_plugin_info
  Get the plugin info
• init_core_context
  Initialize the core context
• init_piper_context
  Initialize the piper context
• init_rag_core_context
  Initialize the core context for RAG scenarios.
• init_sd_context_with_full_model
  Initialize the stable diffusion context with the given full diffusion model
• init_sd_context_with_standalone_model
  Initialize the stable diffusion context with the given standalone diffusion model
• init_whisper_context
  Initialize the whisper context
• running_mode
  Return the current running mode.