LlamaCppSampler + LlamaCppSamplerFactory implementation (v2.3.10 seam impl). More...
#include "llama_cpp_sampler.h"#include <entropic/types/logging.h>#include <llama.h>#include <cstdint>#include <memory>#include <unordered_map>#include <vector>
Go to the source code of this file.
Namespaces | |
| namespace | entropic |
| Activate model on GPU (WARM → ACTIVE). | |
Variables | |
| static auto | logger = entropic::log::get("inference.sampler") |
Detailed Description
LlamaCppSampler + LlamaCppSamplerFactory implementation (v2.3.10 seam impl).
Chain construction logic was lifted verbatim from the legacy LlamaCppBackend::create_sampler. The gates (temperature > 0, repeat_penalty != 1, etc.) and ordering (grammar → penalties → temperature → top-k → top-p → min-p → dist) are unchanged. This file is uncovered by the CPU-only unit-test gate by design — its job is to be a thin pass-through to llama.cpp. Coverage of the LOGIC that consumes Samplers lives in LlamaCppBackend tests (which mock the SamplerFactory).
Definition in file llama_cpp_sampler.cpp.
Variable Documentation
◆ logger
|
static |
Definition at line 31 of file llama_cpp_sampler.cpp.
