Model information. More...

#include <ml_types.h>

Data Fields
char	name [ODP_ML_MODEL_NAME_LEN]
	Model name.

uint64_t	model_version
	Model version number.

uint64_t	interface_version
	Model interface version number.

uint32_t	engine_id
	Engine ID to which the model is assigned.

uint32_t	index
	Model index assigned by the implementation.

uint32_t	num_inputs
	Number of model inputs.

uint32_t	num_outputs
	Number of model outputs.

union {

struct {

uint32_t input_quant_info: 1
	Input quantization information provision. More...

uint32_t output_quant_info: 1
	Output quantization information provision. More...

}
	Auxiliary bit fields.

uint32_t all
	All bits of the bit field structure. More...

}	aux
	Auxiliary information regarding the model and its inputs / outputs.

Detailed Description

Model information.

Examples: odp_ml_perf.c, and odp_ml_run.c.

Definition at line 606 of file api/spec/ml_types.h.

Field Documentation

◆ name

char odp_ml_model_info_t::name[ODP_ML_MODEL_NAME_LEN]

Model name.

Definition at line 608 of file api/spec/ml_types.h.

◆ model_version

uint64_t odp_ml_model_info_t::model_version

Model version number.

Version number of the model binary. The number changes when the model is modified in any way.

Definition at line 616 of file api/spec/ml_types.h.

◆ interface_version

uint64_t odp_ml_model_info_t::interface_version

Model interface version number.

The model interface version number changes only when model input or output data format is modified. Data formats are the same for two model versions that have the same interface version number.

Definition at line 625 of file api/spec/ml_types.h.

◆ engine_id

uint32_t odp_ml_model_info_t::engine_id

Engine ID to which the model is assigned.

Definition at line 628 of file api/spec/ml_types.h.

◆ index

uint32_t odp_ml_model_info_t::index

Model index assigned by the implementation.

Definition at line 631 of file api/spec/ml_types.h.

◆ num_inputs

uint32_t odp_ml_model_info_t::num_inputs

Number of model inputs.

Definition at line 634 of file api/spec/ml_types.h.

◆ num_outputs

uint32_t odp_ml_model_info_t::num_outputs

Number of model outputs.

Definition at line 637 of file api/spec/ml_types.h.

◆ input_quant_info

uint32_t odp_ml_model_info_t::input_quant_info

Input quantization information provision.

When set to 1, model input information provides quantization information. If set, each input needs to be separately checked for information validity (see odp_ml_input_info_t::quant_info).

When set to 0, no quantization information is provided for inputs.

Definition at line 652 of file api/spec/ml_types.h.

◆ output_quant_info

uint32_t odp_ml_model_info_t::output_quant_info

Output quantization information provision.

When set to 1, model output information provides quantization information. If set, each output needs to be separately checked for information validity (see odp_ml_output_info_t::quant_info).

When set to 0, no quantization information is provided for outputs.

Definition at line 663 of file api/spec/ml_types.h.

◆ all

uint32_t odp_ml_model_info_t::all

All bits of the bit field structure.

This field can be used for bitwise operations over the entire structure.

Definition at line 672 of file api/spec/ml_types.h.

The documentation for this struct was generated from the following file:

api/spec/ml_types.h

Data Fields