auto_schedule

Description

Generates a schedule object based on the defined compute process.

Prototype

auto_schedule(outs, option=None)

Parameters

outs: a list of output tensors. Supports one or more outputs.

option: configuration option used for RL operator search. Can be ignored for Auto Schedule.

Returns

schedule: compute schedule of the operator

Applicability

Atlas 200/300/500 Inference Product

Atlas Training Series Product

Example

from tbe import tvm 
from tbe import dsl

shape = (28,28)
dtype = "float16"
# Define the inputs.
data = tvm.placeholder(shape, name="data", dtype=dtype)
# Describe the compute process of the operator.
res = dsl.vabs(data)
with tvm.target.cce():
     # Generate a schedule object.
      sch = dsl.auto_schedule(res)