016eb43741
* [PT FE]: fix issues with squeeze producing dynamic rank * Update test_squeeze.py * Update test_squeeze.py * Update test_slice.py * Apply suggestions from code review
98 lines
2.9 KiB
Python
98 lines
2.9 KiB
Python
# Copyright (C) 2018-2023 Intel Corporation
|
|
# SPDX-License-Identifier: Apache-2.0
|
|
|
|
import numpy as np
|
|
import pytest
|
|
import torch
|
|
from pytorch_layer_test_class import PytorchLayerTest
|
|
|
|
|
|
class TestSlice1D(PytorchLayerTest):
|
|
def _prepare_input(self):
|
|
return (np.array(range(16), np.float32), np.array(self.params, dtype=np.int32))
|
|
|
|
def create_model(self):
|
|
class aten_slice(torch.nn.Module):
|
|
def forward(self, x, params):
|
|
return x[params[0] : params[1] : params[2]]
|
|
|
|
ref_net = None
|
|
|
|
return aten_slice(), ref_net, "aten::slice"
|
|
|
|
@pytest.mark.parametrize(
|
|
"params",
|
|
[[0, -1, 1], [0, -1, 3], [0, 5, 3], [2, 7, 3], [-7, -15, 2], [-1, -7, 2], [5, 2, 1]],
|
|
)
|
|
@pytest.mark.nightly
|
|
@pytest.mark.precommit
|
|
def test_slice1d(self, ie_device, precision, ir_version, params):
|
|
self.params = params
|
|
self._test(
|
|
*self.create_model(),
|
|
ie_device,
|
|
precision,
|
|
ir_version,
|
|
)
|
|
|
|
|
|
class TestSlice2D(PytorchLayerTest):
|
|
def _prepare_input(self):
|
|
return (
|
|
np.array([range(16), range(16, 32)], np.float32),
|
|
np.array(self.params_0a, dtype=np.int32),
|
|
np.array(self.params_1a, dtype=np.int32),
|
|
)
|
|
|
|
def create_model(self):
|
|
class aten_slice(torch.nn.Module):
|
|
def forward(self, x, params_0a, params_1a):
|
|
return x[params_0a[0] : params_0a[1] : params_0a[2], params_1a[0] : params_1a[1] : params_1a[2]]
|
|
|
|
ref_net = None
|
|
|
|
return aten_slice(), ref_net, "aten::slice"
|
|
|
|
@pytest.mark.parametrize(
|
|
"params_0a",
|
|
[[0, -1, 1], [0, -1, 3], [0, 5, 3], [2, 7, 3], [-7, -15, 2], [-1, -7, 2], [5, 2, 1]],
|
|
)
|
|
@pytest.mark.parametrize(
|
|
"params_1a",
|
|
[[0, -1, 1], [0, -1, 3], [0, 5, 3], [2, 7, 3], [-7, -15, 2], [-1, -7, 2], [5, 2, 1]],
|
|
)
|
|
@pytest.mark.nightly
|
|
@pytest.mark.precommit
|
|
def test_slice2d(self, ie_device, precision, ir_version, params_0a, params_1a):
|
|
self.params_0a = params_0a
|
|
self.params_1a = params_1a
|
|
self._test(
|
|
*self.create_model(),
|
|
ie_device,
|
|
precision,
|
|
ir_version,
|
|
)
|
|
|
|
|
|
class TestSliceAndSqueeze(PytorchLayerTest):
|
|
def _prepare_input(self):
|
|
return (np.random.randn(1, 1, 32).astype(np.float32),)
|
|
|
|
def create_model(self):
|
|
class aten_slice(torch.nn.Module):
|
|
def __init__(self) -> None:
|
|
super().__init__()
|
|
|
|
def forward(self, x):
|
|
a = torch.squeeze(x, 1)
|
|
return a[:, None, :]
|
|
|
|
ref_net = None
|
|
|
|
return aten_slice(), ref_net, "aten::slice"
|
|
|
|
@pytest.mark.nightly
|
|
@pytest.mark.precommit
|
|
def test_slice_and_squeeze(self, ie_device, precision, ir_version):
|
|
self._test(*self.create_model(), ie_device, precision, ir_version, dynamic_shapes=False)
|