test_quantized_nn_mods.py

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from __future__ import unicode_literals

import torch
import torch.nn.quantized as nnq
import torch.nn.quantized.dynamic as nnqd
import torch.nn._intrinsic.quantized as nnq_fused
import torch.nn.quantized.functional as qF
from torch.nn.quantized.modules import Conv2d
from torch.nn._intrinsic.quantized import ConvReLU2d
import torch.quantization
from common_utils import run_tests, tempfile
from common_quantization import QuantizationTestCase, no_deadline, prepare_dynamic
from common_quantized import _calculate_dynamic_qparams
from hypothesis import given
from hypothesis import strategies as st
import unittest

'''
Note that tests in this file are just API test, to make sure we wrapped the
quantized operator implementations correctly in the user facing APIs, these are
not correctness test for the underlying quantized operators. For correctness
test please see `caffe2/test/test_quantized.py`.
'''


class FunctionalAPITest(QuantizationTestCase):
    def test_relu_api(self):
        X = torch.arange(-5, 5, dtype=torch.float)
        scale = 2.0
        zero_point = 1
        qX = torch.quantize_linear(X, scale=scale, zero_point=zero_point, dtype=torch.quint8)
        qY = torch.relu(qX)
        qY_hat = qF.relu(qX)
        self.assertEqual(qY, qY_hat)


class DynamicModuleAPITest(QuantizationTestCase):
    @no_deadline
    @unittest.skipIf(
        not torch.fbgemm_is_cpu_supported(),
        " Quantized operations require FBGEMM. FBGEMM is only optimized for CPUs"
        " with instruction set support avx2 or newer.",
    )
    @given(
        batch_size=st.integers(1, 5),
        in_features=st.integers(16, 32),
        out_features=st.integers(4, 8),
        use_bias=st.booleans(),
        use_default_observer=st.booleans(),
    )
    def test_linear_api(self, batch_size, in_features, out_features, use_bias, use_default_observer):
        """test API functionality for nn.quantized.dynamic.Linear"""
        W = torch.rand(out_features, in_features).float()
        W_scale, W_zp = _calculate_dynamic_qparams(W, torch.qint8)
        W_q = torch.quantize_linear(W, W_scale, W_zp, torch.qint8)
        X = torch.rand(batch_size, in_features).float()
        B = torch.rand(out_features).float() if use_bias else None
        qlinear = nnqd.Linear(in_features, out_features)
        # Run module with default-initialized parameters.
        # This tests that the constructor is correct.
        qlinear(X)
        qlinear.set_weight(W_q)

        # Simple round-trip test to ensure weight()/set_weight() API
        self.assertEqual(qlinear.weight(), W_q)
        W_pack = qlinear._packed_weight
        qlinear.bias = B if use_bias else None
        Z_dq = qlinear(X)

        # Check if the module implementation matches calling the
        # ops directly
        Z_ref = torch.ops.quantized.fbgemm_linear_dynamic(X, W_pack, B)
        self.assertEqual(Z_ref, Z_dq)

        # Test serialization of dynamic quantized Linear Module using state_dict
        model_dict = qlinear.state_dict()
        self.assertEqual(model_dict['weight'], W_q)
        if use_bias:
            self.assertEqual(model_dict['bias'], B)
        with tempfile.TemporaryFile() as f:
            torch.save(model_dict, f)
            f.seek(0)
            loaded_dict = torch.load(f)
        for key in model_dict:
            self.assertEqual(model_dict[key], loaded_dict[key])
        loaded_qlinear = nnqd.Linear(in_features, out_features)
        loaded_qlinear.load_state_dict(loaded_dict)

        linear_unpack = torch.ops.quantized.fbgemm_linear_unpack
        self.assertEqual(linear_unpack(qlinear._packed_weight),
                         linear_unpack(loaded_qlinear._packed_weight))
        if use_bias:
            self.assertEqual(qlinear.bias, loaded_qlinear.bias)
        self.assertTrue(dir(qlinear) == dir(loaded_qlinear))
        self.assertTrue(hasattr(qlinear, '_packed_weight'))
        self.assertTrue(hasattr(loaded_qlinear, '_packed_weight'))
        self.assertTrue(hasattr(qlinear, 'weight'))
        self.assertTrue(hasattr(loaded_qlinear, 'weight'))

        self.assertEqual(qlinear.weight(), loaded_qlinear.weight())
        self.assertEqual(qlinear.weight(), torch.ops.quantized.fbgemm_linear_unpack(qlinear._packed_weight))
        Z_dq2 = qlinear(X)
        self.assertEqual(Z_dq, Z_dq2)

        # test serialization of module directly
        with tempfile.TemporaryFile() as f:
            torch.save(qlinear, f)
            f.seek(0)
            loaded = torch.load(f)
        # This check is disabled pending an issue in PyTorch serialization:
        # https://github.com/pytorch/pytorch/issues/24045
        # self.assertEqual(qlinear.weight(), loaded.weight())
        self.assertEqual(qlinear.zero_point, loaded.zero_point)

        # Test JIT
        self.checkScriptable(qlinear, list(zip([X], [Z_ref])), check_save_load=True)

        # Test from_float
        float_linear = torch.nn.Linear(in_features, out_features).float()
        if use_default_observer:
            float_linear.qconfig = torch.quantization.default_dynamic_qconfig
        prepare_dynamic(float_linear)
        float_linear(X.float())
        quantized_float_linear = nnqd.Linear.from_float(float_linear)

        # Smoke test to make sure the module actually runs
        quantized_float_linear(X)

        # Smoke test extra_repr
        str(quantized_float_linear)


class ModuleAPITest(QuantizationTestCase):
    def test_relu(self):
        relu_module = nnq.ReLU()
        relu6_module = nnq.ReLU6()

        x = torch.arange(-10, 10, dtype=torch.float)
        y_ref = torch.relu(x)
        y6_ref = torch.nn.modules.ReLU6()(x)

        qx = torch.quantize_linear(x, 1.0, 0, dtype=torch.qint32)
        qy = relu_module(qx)
        qy6 = relu6_module(qx)

        self.assertEqual(y_ref, qy.dequantize(),
                         message="ReLU module API failed")
        self.assertEqual(y6_ref, qy6.dequantize(),
                         message="ReLU6 module API failed")


    @no_deadline
    @unittest.skipIf(
        not torch.fbgemm_is_cpu_supported(),
        " Quantized operations require FBGEMM. FBGEMM is only optimized for CPUs"
        " with instruction set support avx2 or newer.",
    )
    @given(
        batch_size=st.integers(1, 5),
        in_features=st.integers(16, 32),
        out_features=st.integers(4, 8),
        use_bias=st.booleans(),
        use_fused=st.booleans(),
    )
    def test_linear_api(self, batch_size, in_features, out_features, use_bias, use_fused):
        """test API functionality for nn.quantized.linear and nn._intrinsic.quantized.linear_relu"""
        W = torch.rand(out_features, in_features).float()
        W_q = torch.quantize_linear(W, 0.1, 4, torch.qint8)
        X = torch.rand(batch_size, in_features).float()
        X_q = torch.quantize_linear(X, 0.2, 10, torch.quint8)
        B = torch.rand(out_features).float() if use_bias else None
        B_q = torch.quantize_linear(B, W_q.q_scale() * X_q.q_scale(), 0, torch.qint32) if use_bias else None
        scale = 0.5
        zero_point = 3
        if use_fused:
            qlinear = nnq_fused.LinearReLU(in_features, out_features)
        else:
            qlinear = nnq.Linear(in_features, out_features)

        # Run module with default-initialized parameters.
        # This tests that the constructor is correct.
        qlinear(X_q)

        qlinear.set_weight(W_q)
        # Simple round-trip test to ensure weight()/set_weight() API
        self.assertEqual(qlinear.weight(), W_q)
        W_pack = qlinear._packed_weight
        qlinear.bias = B_q if use_bias else None

        qlinear.scale = float(scale)
        qlinear.zero_point = int(zero_point)
        Z_q = qlinear(X_q)
        # Check if the module implementation matches calling the
        # ops directly
        if use_fused:
            Z_ref = torch.ops.quantized.fbgemm_linear_relu(X_q, W_pack, B_q, scale, zero_point)
        else:
            Z_ref = torch.ops.quantized.fbgemm_linear(X_q, W_pack, B_q, scale, zero_point)
        self.assertEqual(Z_ref, Z_q)

        # Test serialization of quantized Linear Module using state_dict

        model_dict = qlinear.state_dict()
        self.assertEqual(model_dict['weight'], W_q)
        if use_bias:
            self.assertEqual(model_dict['bias'], B_q)
        with tempfile.TemporaryFile() as f:
            torch.save(model_dict, f)
            f.seek(0)
            loaded_dict = torch.load(f)
        for key in model_dict:
            self.assertEqual(model_dict[key], loaded_dict[key])
        if use_fused:
            loaded_qlinear = nnq_fused.LinearReLU(in_features, out_features)
        else:
            loaded_qlinear = nnq.Linear(in_features, out_features)
        loaded_qlinear.load_state_dict(loaded_dict)

        linear_unpack = torch.ops.quantized.fbgemm_linear_unpack
        self.assertEqual(linear_unpack(qlinear._packed_weight),
                         linear_unpack(loaded_qlinear._packed_weight))
        if use_bias:
            self.assertEqual(qlinear.bias, loaded_qlinear.bias)
        self.assertEqual(qlinear.scale, loaded_qlinear.scale)
        self.assertEqual(qlinear.zero_point, loaded_qlinear.zero_point)
        self.assertTrue(dir(qlinear) == dir(loaded_qlinear))
        self.assertTrue(hasattr(qlinear, '_packed_weight'))
        self.assertTrue(hasattr(loaded_qlinear, '_packed_weight'))
        self.assertTrue(hasattr(qlinear, 'weight'))
        self.assertTrue(hasattr(loaded_qlinear, 'weight'))
        self.assertEqual(qlinear.weight(), loaded_qlinear.weight())
        self.assertEqual(qlinear.weight(), torch.ops.quantized.fbgemm_linear_unpack(qlinear._packed_weight))
        Z_q2 = loaded_qlinear(X_q)
        self.assertEqual(Z_q, Z_q2)

        # test serialization of module directly
        with tempfile.TemporaryFile() as f:
            torch.save(qlinear, f)
            f.seek(0)
            loaded = torch.load(f)
        # This check is disabled pending an issue in PyTorch serialization:
        # https://github.com/pytorch/pytorch/issues/24045
        # self.assertEqual(qlinear.weight(), loaded.weight())
        self.assertEqual(qlinear.bias, loaded.bias)
        self.assertEqual(qlinear.scale, loaded.scale)
        self.assertEqual(qlinear.zero_point, loaded.zero_point)

        # Test JIT
        self.checkScriptable(qlinear, list(zip([X_q], [Z_ref])), check_save_load=True)

        # Test from_float
        float_linear = torch.nn.Linear(in_features, out_features).float()
        float_linear.qconfig = torch.quantization.default_qconfig
        torch.quantization.prepare(float_linear)
        float_linear(X.float())
        quantized_float_linear = torch.quantization.convert(float_linear)

        # Smoke test to make sure the module actually runs
        quantized_float_linear(X_q)

        # Smoke test extra_repr
        str(quantized_float_linear)

    def test_quant_dequant_api(self):
        r = torch.tensor([[1., -1.], [1., -1.]], dtype=torch.float)
        scale, zero_point, dtype = 1.0, 2, torch.qint8
        # testing Quantize API
        qr = torch.quantize_linear(r, scale, zero_point, dtype)
        quant_m = nnq.Quantize(scale, zero_point, dtype)
        qr2 = quant_m(r)
        self.assertEqual(qr, qr2)
        # testing Dequantize API
        rqr = qr.dequantize()
        dequant_m = nnq.DeQuantize()
        rqr2 = dequant_m(qr2)
        self.assertEqual(rqr, rqr2)

    @no_deadline
    @unittest.skipIf(
        not torch.fbgemm_is_cpu_supported(),
        " Quantized operations require FBGEMM. FBGEMM is only optimized for CPUs"
        " with instruction set support avx2 or newer.",
    )
    @given(
        use_bias=st.booleans(),
        use_fused=st.booleans(),
    )
    def test_conv_api(self, use_bias, use_fused):
        """Tests the correctness of the conv module.

        The correctness is defined against the functional implementation.
        """

        N, iC, H, W = 10, 10, 10, 3
        oC, g, kH, kW = 16, 1, 3, 3
        scale, zero_point = 1.0 / 255, 128

        X = torch.randn(N, iC, H, W, dtype=torch.float32)
        X = X.permute([0, 2, 3, 1]).contiguous()
        qX = torch.quantize_linear(X, scale=scale, zero_point=128, dtype=torch.quint8)

        w = torch.randn(oC, iC // g, kH, kW, dtype=torch.float32)

        qw = torch.quantize_linear(w, scale=scale, zero_point=0, dtype=torch.qint8)

        b = torch.randn(oC, dtype=torch.float32) if use_bias else None
        qb = torch.quantize_linear(b, scale=1.0 / 1024, zero_point=0, dtype=torch.qint32) if use_bias else None

        if use_fused:
            conv_under_test = ConvReLU2d(in_channels=iC,
                                         out_channels=oC,
                                         kernel_size=(kH, kW),
                                         stride=1,
                                         padding=0,
                                         dilation=1,
                                         groups=g,
                                         bias=use_bias,
                                         padding_mode='zeros')
        else:
            conv_under_test = Conv2d(in_channels=iC,
                                     out_channels=oC,
                                     kernel_size=(kH, kW),
                                     stride=1,
                                     padding=0,
                                     dilation=1,
                                     groups=g,
                                     bias=use_bias,
                                     padding_mode='zeros')
        # Run module with default-initialized parameters.
        # This tests that the constructor is correct.
        conv_under_test(qX)

        conv_under_test.set_weight(qw)
        conv_under_test.bias = qb
        conv_under_test.scale = scale
        conv_under_test.zero_point = zero_point

        # Test members
        self.assertTrue(hasattr(conv_under_test, '_packed_weight'))
        self.assertTrue(hasattr(conv_under_test, 'scale'))
        self.assertTrue(hasattr(conv_under_test, 'zero_point'))

        # Test properties
        self.assertEqual(qw, conv_under_test.weight())
        self.assertEqual(qb, conv_under_test.bias)
        self.assertEqual(scale, conv_under_test.scale)
        self.assertEqual(zero_point, conv_under_test.zero_point)

        # Test forward
        result_under_test = conv_under_test(qX)
        result_reference = qF.conv2d(qX, qw, bias=qb,
                                     scale=scale, zero_point=zero_point,
                                     stride=1, padding=0,
                                     dilation=1, groups=g, dtype=torch.quint8
                                     )
        if use_fused:
            # result_reference < zero_point doesn't work for qtensor yet
            # result_reference[result_reference < zero_point] = zero_point
            MB, OC, OH, OW = result_reference.size()
            for i in range(MB):
                for j in range(OC):
                    for h in range(OH):
                        for w in range(OW):
                            if result_reference[i][j][h][w].int_repr() < zero_point:
                                # assign 0. that gets converted to zero_point
                                result_reference[i][j][h][w] = 0.

        self.assertEqual(result_reference, result_under_test,
                         message="Tensors are not equal.")

        # Test serialization of quantized Conv Module using state_dict
        model_dict = conv_under_test.state_dict()
        self.assertEqual(model_dict['weight'], qw)
        if use_bias:
            self.assertEqual(model_dict['bias'], qb)
        with tempfile.NamedTemporaryFile() as f:
            torch.save(model_dict, f)
            f.seek(0)
            loaded_dict = torch.load(f)
        for key in model_dict:
            self.assertEqual(loaded_dict[key], model_dict[key])
        if use_fused:
            loaded_conv_under_test = ConvReLU2d(in_channels=iC,
                                                out_channels=oC,
                                                kernel_size=(kH, kW),
                                                stride=1,
                                                padding=0,
                                                dilation=1,
                                                groups=g,
                                                bias=use_bias,
                                                padding_mode='zeros')
        else:
            loaded_conv_under_test = Conv2d(in_channels=iC,
                                            out_channels=oC,
                                            kernel_size=(kH, kW),
                                            stride=1,
                                            padding=0,
                                            dilation=1,
                                            groups=g,
                                            bias=use_bias,
                                            padding_mode='zeros')
        loaded_conv_under_test.load_state_dict(loaded_dict)
        self.assertEqual(loaded_conv_under_test.weight(), conv_under_test.weight())
        if use_bias:
            self.assertEqual(loaded_conv_under_test.bias, conv_under_test.bias)
        self.assertEqual(loaded_conv_under_test.scale, conv_under_test.scale)
        self.assertEqual(loaded_conv_under_test.zero_point, conv_under_test.zero_point)
        self.assertTrue(dir(loaded_conv_under_test) == dir(conv_under_test))
        self.assertTrue(hasattr(conv_under_test, '_packed_weight'))
        self.assertTrue(hasattr(loaded_conv_under_test, '_packed_weight'))
        self.assertTrue(hasattr(conv_under_test, 'weight'))
        self.assertTrue(hasattr(loaded_conv_under_test, 'weight'))
        self.assertEqual(loaded_conv_under_test.weight(), conv_under_test.weight())
        self.assertEqual(loaded_conv_under_test.weight(), qw)
        loaded_result = loaded_conv_under_test(qX)
        self.assertEqual(loaded_result, result_reference)

        with tempfile.NamedTemporaryFile() as f:
            torch.save(conv_under_test, f)
            f.seek(0)
            loaded_conv = torch.load(f)

        self.assertEqual(conv_under_test.bias, loaded_conv.bias)
        self.assertEqual(conv_under_test.scale, loaded_conv.scale)
        self.assertEqual(conv_under_test.zero_point, loaded_conv.zero_point)

        # JIT testing
        self.checkScriptable(conv_under_test, list(zip([qX], [result_reference])), check_save_load=True)

        # Test from_float
        float_conv = torch.nn.Conv2d(in_channels=iC,
                                     out_channels=oC,
                                     kernel_size=(kH, kW),
                                     stride=1,
                                     padding=0,
                                     dilation=1,
                                     groups=g,
                                     bias=use_bias,
                                     padding_mode='zeros').float()
        float_conv.qconfig = torch.quantization.default_qconfig
        torch.quantization.prepare(float_conv)
        float_conv(X.float())
        quantized_float_conv = torch.quantization.convert(float_conv)

        # Smoke test to make sure the module actually runs
        quantized_float_conv(qX)
        # Check that bias is quantized based on output scale
        if use_bias:
            qbias = torch.quantize_linear(float_conv.bias, quantized_float_conv.scale / 2**16, 0, torch.qint32)
            self.assertEqual(quantized_float_conv.bias.dequantize(), qbias.dequantize())
        # Smoke test extra_repr
        str(quantized_float_conv)

    def test_pool_api(self):
        """Tests the correctness of the pool module.

        The correctness is defined against the functional implementation.
        """
        N, C, H, W = 10, 10, 10, 3
        kwargs = {
            'kernel_size': 2,
            'stride': None,
            'padding': 0,
            'dilation': 1
        }

        scale, zero_point = 1.0 / 255, 128

        X = torch.randn(N, C, H, W, dtype=torch.float32)
        qX = torch.quantize_linear(X, scale=scale, zero_point=zero_point,
                                   dtype=torch.quint8)
        qX_expect = torch.nn.functional.max_pool2d(qX, **kwargs)

        pool_under_test = torch.nn.quantized.MaxPool2d(**kwargs)
        qX_hat = pool_under_test(qX)
        self.assertEqual(qX_expect, qX_hat)

        # JIT Testing
        self.checkScriptable(pool_under_test, list(zip([X], [qX_expect])))

if __name__ == '__main__':
    run_tests()