Module model_lifecycle 

Catalog LLM download → GGUF shard map → WAL → lifecycle (in-process).

Structs

ActiveModelRecord

InstallManifest

ModelInstallResult

ModelStatus

VaultDuplicateGroup

Exact duplicate GGUF files, grouped by content rather than filename.

VaultGgufEntry

Discovered weight container in a vault directory (CLI / lifecycle tooling).

Enums

ModelError

Constants

MEMORY_FLOOR_MB

UI RAM ceiling for LLM arena pressure (matches 0.0.12 Flutter plan).

Functions

activate_model

activate_model_for_id

activate_vault_gguf

Map and activate a vault weight file (.p64 preferred, or .gguf) without catalog install.

audit_vault_duplicates

Hash only same-sized GGUF candidates and report byte-identical duplicates.

finalize_llm_install

finalize_local_gguf

Register and activate a user-selected .gguf file at an arbitrary path.

get_kv_cache_used_mb

get_last_decode_tok_s

Last measured tok/s from a completed turn, or None if never measured this process.

get_last_decode_tok_s_at_unix

get_llm_memory_bytes

get_model_lifecycle_state

get_model_status

get_thermal_state_label

install_catalog_llm

lifecycle_label

load_install_manifest

models_dir

record_kv_cache_used_mb

record_last_decode_tok_s

Record measured decode/chat throughput from a completed generation turn.

record_llm_memory_bytes

record_llm_memory_sample

resolve_vault_model

Resolve a model reference (filename, stem, or path) inside vault_dir.

scan_vault_gguf

Recursively scan vault_dir for .p64 and .gguf weight containers.

task_orchestrator

Shared orchestrator for chat inference and model lifecycle transitions.

unload_active_model

wait_for_eviction_scrub

Block until eviction scrub completes or timeout elapses.