Skip to content

Generalization for mac estimation for conv2d when groups is not 1 #222

Open
@poyupaulchen

Description

@poyupaulchen


if groups of conv2d is considered, macs should be like this for generalization

macs = (
                        filter["shape"][1]
                        * filter["shape"][2]
                        * filter["shape"][3] 
                        * outputs["shape"][1]
                        * outputs["shape"][2]
                        * outputs["shape"][3] 
                    )

The total # of MACs will be sizeof(filter) * sizeof(output)

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions