Documentation

memory

Memory benchmark result models.

Classes

class

CorpusStats

(BaseModel)

Corpus statistics.

Attributes

NameTypeDescription
edge_featuresint—
max_nodeint—
max_slotint—
namestr—
node_featuresint—
node_typesint—

Methods

to_dict(self) → dict

Convert to JSON-serializable dictionary.

class

MemoryBenchmarkResult

(BaseModel)

Aggregated memory benchmark results for one corpus.

Attributes

NameTypeDescription
cf_fork_statsStatisticalSummary | None—
cf_load_time_statsStatisticalSummary | None—
cf_memory_statsStatisticalSummary | None—
cf_spawn_statsStatisticalSummary | None—
corpusstr—
corpus_statsCorpusStats—
measurementslist[MemoryMeasurement]—
tf_fork_statsStatisticalSummary | None—
tf_load_time_statsStatisticalSummary | None—
tf_memory_statsStatisticalSummary | None—
tf_spawn_statsStatisticalSummary | None—

Methods

get_measurements_by_impl(self, implementation: Literal[('TF', 'CF')]) → list[MemoryMeasurement]

Get measurements for a specific implementation.

Parameters
  • implementation: Literal[('TF', 'CF')]
get_measurements_by_mode(self, mode: Literal[('single', 'spawn', 'fork')]) → list[MemoryMeasurement]

Get measurements for a specific mode.

Parameters
  • mode: Literal[('single', 'spawn', 'fork')]
to_dict(self) → dict

Convert to JSON-serializable dictionary.

to_json(self) → str

Serialize to JSON string.

class

MemoryMeasurement

(BaseModel)

Single memory measurement.

Attributes

NameTypeDescription
cache_size_mbfloat | None—
compile_time_sfloat | None—
corpusstr—
implementationLiteral[('TF', 'CF')]—
load_time_sfloat—
memory_used_mbfloatCalculate memory used (after - before).
modeLiteral[('single', 'spawn', 'fork')]—
num_workersint | None—
per_worker_rss_mbfloat | None—
rss_after_mbfloat—
rss_before_mbfloat—
run_idint—
total_rss_mbfloat | None—

Methods

to_dict(self) → dict

Convert to JSON-serializable dictionary.