def baseline_mwe(nO, nP, depth):
from thinc.neural._classes.model import Model
from thinc.neural._classes.resnet import Residual
from thinc.neural._classes.convolution import ExtractWindow
from thinc.neural._classes.layernorm import LayerNorm
from thinc.api import chain, clone, with_flatten
maxout = Maxout(nO, nO*3, pieces=nP)
normalize = LayerNorm(maxout)
with Model.define_operators({'>>': chain, '**': clone}):
model = Residual(ExtractWindow(nW=1) >> normalize)
model = with_flatten(chain(*([model]*depth)))
model.maxout = maxout
model.normalize = normalize
return model
def test_check_operator_is_defined_passes(model, dummy, operator):
checker = check.operator_is_defined(operator)
checked = checker(dummy)
with Model.define_operators({"+": None}):
checked(model, None)
