Python torch.randn() Examples

def init_duvenaud(self, params):
        learn_args = []
        learn_modules = []
        args = {}

        # Filter degree 0 (the message will be 0 and therefore there is no update
        args['deg'] = [i for i in params['deg'] if i!=0]
        args['in'] = params['in']
        args['out'] = params['out']

        # Define a parameter matrix H for each degree.
        learn_args.append(torch.nn.Parameter(torch.randn(len(args['deg']), args['in'], args['out'])))

        return nn.ParameterList(learn_args), nn.ModuleList(learn_modules), args

    # GG-NN, Li et al. 

def forward(self, x):
        out = self.pre_layers(x)
        out = self.a3(out)
        out = self.b3(out)
        out = self.maxpool(out)
        out = self.a4(out)
        out = self.b4(out)
        out = self.c4(out)
        out = self.d4(out)
        out = self.e4(out)
        out = self.maxpool(out)
        out = self.a5(out)
        out = self.b5(out)
        out = self.avgpool(out)
        out = out.view(out.size(0), -1)
        out = self.linear(out)
        return out

# net = GoogLeNet()
# x = torch.randn(1,3,32,32)
# y = net(Variable(x))
# print(y.size()) 

def _make_layers(self, cfg):
        layers = []
        in_channels = 3
        for x in cfg:
            if x == 'M':
                layers += [nn.MaxPool2d(kernel_size=2, stride=2)]
            else:
                layers += [nn.Conv2d(in_channels, x, kernel_size=3, padding=1),
                           nn.BatchNorm2d(x),
                           nn.ReLU(inplace=True)]
                in_channels = x
        layers += [nn.AvgPool2d(kernel_size=1, stride=1)]
        return nn.Sequential(*layers)

# net = VGG('VGG11')
# x = torch.randn(2,3,32,32)
# print(net(Variable(x)).size()) 

def init_duvenaud(self, params):
        learn_args = []
        learn_modules = []
        args = {}

        args['out'] = params['out']

        # Define a parameter matrix W for each layer.
        for l in range(params['layers']):
            learn_args.append(nn.Parameter(torch.randn(params['in'][l], params['out'])))

        # learn_modules.append(nn.Linear(params['out'], params['target']))

        learn_modules.append(NNet(n_in=params['out'], n_out=params['target']))
        return nn.ParameterList(learn_args), nn.ModuleList(learn_modules), args

    # GG-NN, Li et al. 

def _make_layers(self, cfg):
        layers = []
        in_channels = 3
        for x in cfg:
            if x == 'M':
                layers += [nn.MaxPool2d(kernel_size=2, stride=2)]
            else:
                layers += [nn.Conv2d(in_channels, x, kernel_size=3, padding=1),
                           nn.BatchNorm2d(x),
                           nn.ReLU(inplace=True)]
                in_channels = x
        layers += [nn.AvgPool2d(kernel_size=1, stride=1)]
        return nn.Sequential(*layers)

# net = VGG('VGG11')
# x = torch.randn(2,3,32,32)
# print(net(Variable(x)).size()) 

def test_res2net_backbone():
    with pytest.raises(KeyError):
        # Res2Net depth should be in [50, 101, 152]
        Res2Net(depth=18)

    # Test Res2Net with scales 4, base_width 26
    model = Res2Net(depth=50, scales=4, base_width=26)
    for m in model.modules():
        if is_block(m):
            assert m.scales == 4
    model.init_weights()
    model.train()

    imgs = torch.randn(1, 3, 224, 224)
    feat = model(imgs)
    assert len(feat) == 4
    assert feat[0].shape == torch.Size([1, 256, 56, 56])
    assert feat[1].shape == torch.Size([1, 512, 28, 28])
    assert feat[2].shape == torch.Size([1, 1024, 14, 14])
    assert feat[3].shape == torch.Size([1, 2048, 7, 7]) 

def test_batch_InverseMelScale(self):
        n_mels = 32
        n_stft = 5
        mel_spec = torch.randn(2, n_mels, 32) ** 2

        # Single then transform then batch
        expected = torchaudio.transforms.InverseMelScale(n_stft, n_mels)(mel_spec).repeat(3, 1, 1, 1)

        # Batch then transform
        computed = torchaudio.transforms.InverseMelScale(n_stft, n_mels)(mel_spec.repeat(3, 1, 1, 1))

        # shape = (3, 2, n_mels, 32)

        # Because InverseMelScale runs SGD on randomly initialized values so they do not yield
        # exactly same result. For this reason, tolerance is very relaxed here.
        self.assertEqual(computed, expected, atol=1.0, rtol=1e-5) 

def test_resnext_backbone():
    with pytest.raises(KeyError):
        # ResNeXt depth should be in [50, 101, 152]
        ResNeXt(depth=18)

    # Test ResNeXt with group 32, base_width 4
    model = ResNeXt(depth=50, groups=32, base_width=4)
    for m in model.modules():
        if is_block(m):
            assert m.conv2.groups == 32
    model.init_weights()
    model.train()

    imgs = torch.randn(1, 3, 224, 224)
    feat = model(imgs)
    assert len(feat) == 4
    assert feat[0].shape == torch.Size([1, 256, 56, 56])
    assert feat[1].shape == torch.Size([1, 512, 28, 28])
    assert feat[2].shape == torch.Size([1, 1024, 14, 14])
    assert feat[3].shape == torch.Size([1, 2048, 7, 7]) 

def _make_layers(self, cfg):
        layers = []
        in_channels = 3
        for x in cfg:
            if x == 'M':
                layers += [nn.MaxPool2d(kernel_size=2, stride=2)]
            else:
                layers += [nn.Conv2d(in_channels, x, kernel_size=3, padding=1),
                           nn.BatchNorm2d(x),
                           nn.ReLU(inplace=True)]
                in_channels = x
        layers += [nn.AvgPool2d(kernel_size=1, stride=1)]
        return nn.Sequential(*layers)

# net = VGG('VGG11')
# x = torch.randn(2,3,32,32)
# print(net(Variable(x)).size()) 

def forward(self, x):
        out = self.pre_layers(x)
        out = self.a3(out)
        out = self.b3(out)
        out = self.maxpool(out)
        out = self.a4(out)
        out = self.b4(out)
        out = self.c4(out)
        out = self.d4(out)
        out = self.e4(out)
        out = self.maxpool(out)
        out = self.a5(out)
        out = self.b5(out)
        out = self.avgpool(out)
        out = out.view(out.size(0), -1)
        out = self.linear(out)
        return out

# net = GoogLeNet()
# x = torch.randn(1,3,32,32)
# y = net(Variable(x))
# print(y.size()) 

def forward_dummy(self, img):
        """Used for computing network flops.

        See `mmdetection/tools/get_flops.py`
        """
        outs = ()
        # backbone
        x = self.extract_feat(img)
        # rpn
        if self.with_rpn:
            rpn_outs = self.rpn_head(x)
            outs = outs + (rpn_outs, )
        proposals = torch.randn(1000, 4).to(img.device)
        # roi_head
        roi_outs = self.roi_head.forward_dummy(x, proposals)
        outs = outs + (roi_outs, )
        return outs 

def translate(self, tensors, cond, num_decode, enum_root):
        assert enum_root 
        tensors = make_cuda(tensors)
        root_vecs, tree_vecs, graph_vecs = self.encode(tensors)

        cond = cond.view(1,1,-1)
        tree_cond = cond.expand(num_decode, tree_vecs.size(1), -1)
        graph_cond = cond.expand(num_decode, graph_vecs.size(1), -1)

        if enum_root:
            repeat = num_decode // len(root_vecs)
            modulo = num_decode % len(root_vecs)
            root_vecs = torch.cat([root_vecs] * repeat + [root_vecs[:modulo]], dim=0)
            tree_vecs = torch.cat([tree_vecs] * repeat + [tree_vecs[:modulo]], dim=0)
            graph_vecs = torch.cat([graph_vecs] * repeat + [graph_vecs[:modulo]], dim=0)

        z_tree = torch.randn(num_decode, 1, self.latent_size).expand(-1, tree_vecs.size(1), -1).cuda()
        z_graph = torch.randn(num_decode, 1, self.latent_size).expand(-1, graph_vecs.size(1), -1).cuda()
        z_tree_vecs = self.W_tree( torch.cat([tree_vecs, z_tree, tree_cond], dim=-1) )
        z_graph_vecs = self.W_graph( torch.cat([graph_vecs, z_graph, graph_cond], dim=-1) )
        return self.decoder.decode( (root_vecs, z_tree_vecs, z_graph_vecs) ) 

def translate(self, tensors, num_decode, enum_root, greedy=True):
        tensors = make_cuda(tensors)
        root_vecs, tree_vecs, graph_vecs = self.encode(tensors)
        all_smiles = []
        if enum_root:
            repeat = num_decode // len(root_vecs)
            modulo = num_decode % len(root_vecs)
            root_vecs = torch.cat([root_vecs] * repeat + [root_vecs[:modulo]], dim=0)
            tree_vecs = torch.cat([tree_vecs] * repeat + [tree_vecs[:modulo]], dim=0)
            graph_vecs = torch.cat([graph_vecs] * repeat + [graph_vecs[:modulo]], dim=0)
        
        batch_size = len(root_vecs)
        z_tree = torch.randn(batch_size, 1, self.latent_size).expand(-1, tree_vecs.size(1), -1).cuda()
        z_graph = torch.randn(batch_size, 1, self.latent_size).expand(-1, graph_vecs.size(1), -1).cuda()
        z_tree_vecs = self.W_tree( torch.cat([tree_vecs, z_tree], dim=-1) )
        z_graph_vecs = self.W_graph( torch.cat([graph_vecs, z_graph], dim=-1) )
        return self.decoder.decode( (root_vecs, z_tree_vecs, z_graph_vecs), greedy=greedy) 

def translate(self, tensors, cond, num_decode, enum_root):
        assert enum_root 
        tensors = make_cuda(tensors)
        root_vecs, tree_vecs, graph_vecs = self.encode(tensors)

        cond = cond.view(1,1,-1)
        tree_cond = cond.expand(num_decode, tree_vecs.size(1), -1)
        graph_cond = cond.expand(num_decode, graph_vecs.size(1), -1)

        if enum_root:
            repeat = num_decode // len(root_vecs)
            modulo = num_decode % len(root_vecs)
            root_vecs = torch.cat([root_vecs] * repeat + [root_vecs[:modulo]], dim=0)
            tree_vecs = torch.cat([tree_vecs] * repeat + [tree_vecs[:modulo]], dim=0)
            graph_vecs = torch.cat([graph_vecs] * repeat + [graph_vecs[:modulo]], dim=0)

        z_tree = torch.randn(num_decode, 1, self.latent_size).expand(-1, tree_vecs.size(1), -1).cuda()
        z_graph = torch.randn(num_decode, 1, self.latent_size).expand(-1, graph_vecs.size(1), -1).cuda()
        z_tree_vecs = self.W_tree( torch.cat([tree_vecs, z_tree, tree_cond], dim=-1) )
        z_graph_vecs = self.W_graph( torch.cat([graph_vecs, z_graph, graph_cond], dim=-1) )
        return self.decoder.decode( (root_vecs, z_tree_vecs, z_graph_vecs) ) 

def translate(self, tensors, num_decode, enum_root, greedy=True):
        tensors = make_cuda(tensors)
        root_vecs, tree_vecs, graph_vecs = self.encode(tensors)
        all_smiles = []
        if enum_root:
            repeat = num_decode // len(root_vecs)
            modulo = num_decode % len(root_vecs)
            root_vecs = torch.cat([root_vecs] * repeat + [root_vecs[:modulo]], dim=0)
            tree_vecs = torch.cat([tree_vecs] * repeat + [tree_vecs[:modulo]], dim=0)
            graph_vecs = torch.cat([graph_vecs] * repeat + [graph_vecs[:modulo]], dim=0)
        
        batch_size = len(root_vecs)
        z_tree = torch.randn(batch_size, 1, self.latent_size).expand(-1, tree_vecs.size(1), -1).cuda()
        z_graph = torch.randn(batch_size, 1, self.latent_size).expand(-1, graph_vecs.size(1), -1).cuda()
        z_tree_vecs = self.W_tree( torch.cat([tree_vecs, z_tree], dim=-1) )
        z_graph_vecs = self.W_graph( torch.cat([graph_vecs, z_graph], dim=-1) )
        return self.decoder.decode( (root_vecs, z_tree_vecs, z_graph_vecs), greedy=greedy) 

def forward(self, x):
        out = self.pre_layers(x)
        out = self.a3(out)
        out = self.b3(out)
        out = self.maxpool(out)
        out = self.a4(out)
        out = self.b4(out)
        out = self.c4(out)
        out = self.d4(out)
        out = self.e4(out)
        out = self.maxpool(out)
        out = self.a5(out)
        out = self.b5(out)
        out = self.avgpool(out)
        out = out.view(out.size(0), -1)
        out = self.linear(out)
        return out

# net = GoogLeNet()
# x = torch.randn(1,3,32,32)
# y = net(Variable(x))
# print(y.size()) 

def dropout(X, drop_prob):
    X = X.float()
    assert 0 <= drop_prob <= 1
    keep_prob = 1 - drop_prob
    if keep_prob == 0:
        return torch.zeros_like(X)
    mask = (torch.randn(X.shape) < keep_prob).float()

    return mask * X / keep_prob 

def test():
    net = ResNet18()
    y = net(Variable(torch.randn(1,3,32,32)))
    print(y.size())

# test() 

def init_params():
    w = torch.randn((num_inputs, 1), requires_grad=True)
    b = torch.zeros(1, requires_grad=True)
    return [w, b] 

def test_batch_Resample(self):
        waveform = torch.randn(2, 2786)

        # Single then transform then batch
        expected = torchaudio.transforms.Resample()(waveform).repeat(3, 1, 1)

        # Batch then transform
        computed = torchaudio.transforms.Resample()(waveform.repeat(3, 1, 1))

        self.assertEqual(computed, expected) 

def test_sliding_window_cmn(self):
        """sliding_window_cmn should be numerically compatible with apply-cmvn-sliding"""
        kwargs = {
            'cmn_window': 600,
            'min_cmn_window': 100,
            'center': False,
            'norm_vars': False,
        }

        tensor = torch.randn(40, 10, dtype=self.dtype, device=self.device)
        result = F.sliding_window_cmn(tensor, **kwargs)
        command = ['apply-cmvn-sliding'] + _convert_args(**kwargs) + ['ark:-', 'ark:-']
        kaldi_result = _run_kaldi(command, 'ark', tensor)
        self.assert_equal(result, expected=kaldi_result) 

def test_batch_MelScale(self):
        specgram = torch.randn(2, 31, 2786)

        # Single then transform then batch
        expected = torchaudio.transforms.MelScale()(specgram).repeat(3, 1, 1, 1)

        # Batch then transform
        computed = torchaudio.transforms.MelScale()(specgram.repeat(3, 1, 1, 1))

        # shape = (3, 2, 201, 1394)
        self.assertEqual(computed, expected) 

def test_compute_deltas(self):
        channel = 13
        n_mfcc = channel * 3
        time = 1021
        win_length = 2 * 7 + 1
        specgram = torch.randn(channel, n_mfcc, time)
        transform = transforms.ComputeDeltas(win_length=win_length)
        computed = transform(specgram)
        self.assertTrue(computed.shape == specgram.shape, (computed.shape, specgram.shape)) 

def init_ggnn(self, params):
        learn_args = []
        learn_modules = []
        args = {}

        args['e_label'] = params['e_label']
        args['in'] = params['in']
        args['out'] = params['out']

        # Define a parameter matrix A for each edge label.
        learn_args.append(nn.Parameter(torch.randn(len(params['e_label']), params['in'], params['out'])))

        return nn.ParameterList(learn_args), nn.ModuleList(learn_modules), args

    # Battaglia et al. (2016), Interaction Networks 

def test():
    net = ShuffleNetG2()
    x = Variable(torch.randn(1,3,32,32))
    y = net(x)
    print(y)

# test() 

def test_densenet():
    net = densenet_cifar()
    x = torch.randn(1,3,32,32)
    y = net(Variable(x))
    print(y)

# test_densenet() 

def main():
    config = json.load(open('../../configs/dcgan_exp_0.json'))
    config = edict(config)
    inp  = torch.autograd.Variable(torch.randn(config.batch_size, config.g_input_size, 1, 1))
    print (inp.shape)
    netD = Generator(config)
    out = netD(inp)
    print (out.shape) 

def test():
    net = PreActResNet18()
    y = net(Variable(torch.randn(1,3,32,32)))
    print(y.size())

# test() 

def test():
    net = SENet18()
    y = net(Variable(torch.randn(1,3,32,32)))
    print(y.size())

# test() 

def test():
    net = DPN92()
    x = Variable(torch.randn(1,3,32,32))
    y = net(x)
    print(y)

# test()