You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
126 lines
3.5 KiB
126 lines
3.5 KiB
import numpy
|
|
from hypothesis.extra.numpy import arrays
|
|
from hypothesis.strategies import floats, integers, just, tuples
|
|
|
|
from thinc.api import Linear, NumpyOps
|
|
|
|
|
|
def get_ops():
|
|
return NumpyOps()
|
|
|
|
|
|
def get_model(W_values, b_values):
|
|
model = Linear(W_values.shape[0], W_values.shape[1], ops=NumpyOps())
|
|
model.initialize()
|
|
model.set_param("W", W_values)
|
|
model.set_param("b", b_values)
|
|
return model
|
|
|
|
|
|
def get_output(input_, W_values, b_values):
|
|
return numpy.einsum("oi,bi->bo", W_values, input_) + b_values
|
|
|
|
|
|
def get_input(nr_batch, nr_in):
|
|
ops = NumpyOps()
|
|
return ops.alloc2f(nr_batch, nr_in)
|
|
|
|
|
|
def lengths(lo=1, hi=10):
|
|
return integers(min_value=lo, max_value=hi)
|
|
|
|
|
|
def shapes(min_rows=1, max_rows=100, min_cols=1, max_cols=100):
|
|
return tuples(lengths(lo=min_rows, hi=max_rows), lengths(lo=min_cols, hi=max_cols))
|
|
|
|
|
|
def ndarrays_of_shape(shape, lo=-10.0, hi=10.0, dtype="float32", width=32):
|
|
if dtype.startswith("float"):
|
|
return arrays(
|
|
dtype, shape=shape, elements=floats(min_value=lo, max_value=hi, width=width)
|
|
)
|
|
else:
|
|
return arrays(dtype, shape=shape, elements=integers(min_value=lo, max_value=hi))
|
|
|
|
|
|
def ndarrays(min_len=0, max_len=10, min_val=-10.0, max_val=10.0):
|
|
return lengths(lo=1, hi=2).flatmap(
|
|
lambda n: ndarrays_of_shape(n, lo=min_val, hi=max_val)
|
|
)
|
|
|
|
|
|
def arrays_BI(min_B=1, max_B=10, min_I=1, max_I=100):
|
|
shapes = tuples(lengths(lo=min_B, hi=max_B), lengths(lo=min_I, hi=max_I))
|
|
return shapes.flatmap(ndarrays_of_shape)
|
|
|
|
|
|
def arrays_BOP(min_B=1, max_B=10, min_O=1, max_O=100, min_P=1, max_P=5):
|
|
shapes = tuples(
|
|
lengths(lo=min_B, hi=max_B),
|
|
lengths(lo=min_O, hi=max_O),
|
|
lengths(lo=min_P, hi=max_P),
|
|
)
|
|
return shapes.flatmap(ndarrays_of_shape)
|
|
|
|
|
|
def arrays_BOP_BO(min_B=1, max_B=10, min_O=1, max_O=100, min_P=1, max_P=5):
|
|
shapes = tuples(
|
|
lengths(lo=min_B, hi=max_B),
|
|
lengths(lo=min_O, hi=max_O),
|
|
lengths(lo=min_P, hi=max_P),
|
|
)
|
|
return shapes.flatmap(
|
|
lambda BOP: tuples(ndarrays_of_shape(BOP), ndarrays_of_shape(BOP[:-1]))
|
|
)
|
|
|
|
|
|
def arrays_BI_BO(min_B=1, max_B=10, min_I=1, max_I=100, min_O=1, max_O=100):
|
|
shapes = tuples(
|
|
lengths(lo=min_B, hi=max_B),
|
|
lengths(lo=min_I, hi=max_I),
|
|
lengths(lo=min_O, hi=max_O),
|
|
)
|
|
return shapes.flatmap(
|
|
lambda BIO: tuples(
|
|
ndarrays_of_shape((BIO[0], BIO[1])), ndarrays_of_shape((BIO[0], BIO[2]))
|
|
)
|
|
)
|
|
|
|
|
|
def arrays_OI_O_BI(
|
|
min_batch=1, max_batch=16, min_out=1, max_out=16, min_in=1, max_in=16
|
|
):
|
|
shapes = tuples(
|
|
lengths(lo=min_batch, hi=max_batch),
|
|
lengths(lo=min_in, hi=max_out),
|
|
lengths(lo=min_in, hi=max_in),
|
|
)
|
|
|
|
def W_b_inputs(shape):
|
|
batch_size, nr_out, nr_in = shape
|
|
W = ndarrays_of_shape((nr_out, nr_in))
|
|
b = ndarrays_of_shape((nr_out,))
|
|
input_ = ndarrays_of_shape((batch_size, nr_in))
|
|
return tuples(W, b, input_)
|
|
|
|
return shapes.flatmap(W_b_inputs)
|
|
|
|
|
|
def arrays_OPFI_BI_lengths(max_B=5, max_P=3, max_F=5, max_I=8):
|
|
shapes = tuples(
|
|
lengths(hi=max_B),
|
|
lengths(hi=max_P),
|
|
lengths(hi=max_F),
|
|
lengths(hi=max_I),
|
|
arrays("int32", shape=(5,), elements=integers(min_value=1, max_value=10)),
|
|
)
|
|
|
|
strat = shapes.flatmap(
|
|
lambda opfi_lengths: tuples(
|
|
ndarrays_of_shape(opfi_lengths[:-1]),
|
|
ndarrays_of_shape((sum(opfi_lengths[-1]), opfi_lengths[-2])),
|
|
just(opfi_lengths[-1]),
|
|
)
|
|
)
|
|
return strat
|