Result

Tuning result of auto_scheduler dialect.

class lorien.dialect.tvm_dial.auto_scheduler_dial.result.AutoSchedulerRecords(target_key: str, workload_key: Optional[str] = None)

The container to maintain the records of an auto_scheduler task.

static decode(record_str: str) Tuple[tvm.auto_scheduler.measure.MeasureInput, tvm.auto_scheduler.measure.MeasureResult]

Decode a string to a record.

static encode(record: Tuple[tvm.auto_scheduler.measure.MeasureInput, tvm.auto_scheduler.measure.MeasureResult]) str

Encode a record to a string.

static gen_record_item(record: Tuple[tvm.auto_scheduler.measure.MeasureInput, tvm.auto_scheduler.measure.MeasureResult])

Generate an item for a record that can be appended to the task item.

gen_task_item() Dict[str, Any]

No additional attribute is required for auto_scheduler.

peak() Tuple[tvm.auto_scheduler.measure.MeasureInput, tvm.auto_scheduler.measure.MeasureResult]

Peak the first record.

pop() Tuple[tvm.auto_scheduler.measure.MeasureInput, tvm.auto_scheduler.measure.MeasureResult]

Pop the worst record in the container and remove the cost.

push(record: Tuple[tvm.auto_scheduler.measure.MeasureInput, tvm.auto_scheduler.measure.MeasureResult])

Push a new record.

Parameters

record (Any) -- The record to be pushed.

to_list(nbest: int = - 1) List[Tuple[tvm.auto_scheduler.measure.MeasureInput, tvm.auto_scheduler.measure.MeasureResult]]

Sort the record (of any layout) to be a list and return the best N.

Parameters

nbest (int) -- The best N records to be returned. Default to return all.

Returns

records -- The sorted list of records.

Return type

List[Tuple[MeasureInput, MeasureResult]]

class lorien.dialect.tvm_dial.auto_scheduler_dial.result.AutoSchedulerTuneResult

The result of a tuning job.

static create_records_by_workloads(log_file_path: str, nbest: int, workload: Optional[lorien.workload.Workload] = None) Sequence[lorien.dialect.tvm_dial.auto_scheduler_dial.result.AutoSchedulerRecords]

Parse records from the tuning log and group them by workloads.

Parameters
  • log_file_path (str) -- The log file path.

  • nbest (int) -- The maximum number of best records to be kept.

  • workload (Optional[Workload]) -- The target workload. If presented, the returnede map will only have one entry.

Returns

records -- A list of created records.

Return type

Sequence[AutoSchedulerRecords]

static gen_features(log_file_path: str, out_path: str)

Featurize tuning logs to be input features of the performance cost model.

Parameters
  • log_file_path (str) -- The log file path. It can be a file path for a single file, or a directory of several log files.

  • out_path (str) -- The path to write generated features and artifacts.