Join the 1-day Testing & QA Summit featuring 15+ Expert Speakers.Register for FREE! Join TestMu Conference

How to use patch_config method in yandex-tank

Best Python code snippet using yandex-tank

xception.py

Source:xception.py

...8    aligned_xception.preprocess_input)9def AlignedXception41Stride16(*args, **kwargs):10    base = AlignedXception41(*args, **kwargs)11    conf = base.get_config()12    conf = patch_config(conf, ['exit_flow/block1/unit1/sepconv3_depthwise'], 'strides', (1, 1))13    conf = patch_config(conf, ['exit_flow/block1/unit1/shortcut'], 'strides', (1, 1))14    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv1_depthwise'], 'dilation_rate', (2, 2))15    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv2_depthwise'], 'dilation_rate', (2, 2))16    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv3_depthwise'], 'dilation_rate', (2, 2))17    patch = models.Model.from_config(conf)18    patch.set_weights(base.get_weights())19    return patch20def AlignedXception41Stride8(*args, **kwargs):21    base = AlignedXception41(*args, **kwargs)22    conf = base.get_config()23    conf = patch_config(conf, ['entry_flow/block3/unit1/sepconv3_depthwise'], 'strides', (1, 1))24    conf = patch_config(conf, ['entry_flow/block3/unit1/shortcut'], 'strides', (1, 1))25    for i in range(8):26        conf = patch_config(conf, [27            'middle_flow/block1/unit{}/sepconv1_depthwise'.format(i + 1)], 'dilation_rate', (2, 2))28        conf = patch_config(conf, [29            'middle_flow/block1/unit{}/sepconv2_depthwise'.format(i + 1)], 'dilation_rate', (2, 2))30        conf = patch_config(conf, [31            'middle_flow/block1/unit{}/sepconv3_depthwise'.format(i + 1)], 'dilation_rate', (2, 2))32    conf = patch_config(conf, ['exit_flow/block1/unit1/sepconv3_depthwise'], 'strides', (1, 1))33    conf = patch_config(conf, ['exit_flow/block1/unit1/shortcut'], 'strides', (1, 1))34    # Output shape is different when using exit dilation rates like in35    # https://github.com/bonlime/keras-deeplab-v3-plus/blob/master/model.py#L26536    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv1_depthwise'], 'dilation_rate', (2, 2))37    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv2_depthwise'], 'dilation_rate', (2, 2))38    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv3_depthwise'], 'dilation_rate', (2, 2))39    patch = models.Model.from_config(conf)40    patch.set_weights(base.get_weights())41    return patch42AlignedXception65 = partial(43    wrap_bone,44    aligned_xception.Xception65,45    aligned_xception.preprocess_input)46def AlignedXception65Stride16(*args, **kwargs):47    base = AlignedXception65(*args, **kwargs)48    conf = base.get_config()49    conf = patch_config(conf, ['exit_flow/block1/unit1/sepconv3_depthwise'], 'strides', (1, 1))50    conf = patch_config(conf, ['exit_flow/block1/unit1/shortcut'], 'strides', (1, 1))51    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv1_depthwise'], 'dilation_rate', (2, 2))52    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv2_depthwise'], 'dilation_rate', (2, 2))53    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv3_depthwise'], 'dilation_rate', (2, 2))54    patch = models.Model.from_config(conf)55    patch.set_weights(base.get_weights())56    return patch57def AlignedXception65Stride8(*args, **kwargs):58    base = AlignedXception65(*args, **kwargs)59    conf = base.get_config()60    conf = patch_config(conf, ['entry_flow/block3/unit1/sepconv3_depthwise'], 'strides', (1, 1))61    conf = patch_config(conf, ['entry_flow/block3/unit1/shortcut'], 'strides', (1, 1))62    for i in range(16):63        conf = patch_config(conf, [64            'middle_flow/block1/unit{}/sepconv1_depthwise'.format(i + 1)], 'dilation_rate', (2, 2))65        conf = patch_config(conf, [66            'middle_flow/block1/unit{}/sepconv2_depthwise'.format(i + 1)], 'dilation_rate', (2, 2))67        conf = patch_config(conf, [68            'middle_flow/block1/unit{}/sepconv3_depthwise'.format(i + 1)], 'dilation_rate', (2, 2))69    conf = patch_config(conf, ['exit_flow/block1/unit1/sepconv3_depthwise'], 'strides', (1, 1))70    conf = patch_config(conf, ['exit_flow/block1/unit1/shortcut'], 'strides', (1, 1))71    # Output shape is different when using exit dilation rates like in72    # https://github.com/bonlime/keras-deeplab-v3-plus/blob/master/model.py#L26573    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv1_depthwise'], 'dilation_rate', (2, 2))74    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv2_depthwise'], 'dilation_rate', (2, 2))75    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv3_depthwise'], 'dilation_rate', (2, 2))76    patch = models.Model.from_config(conf)77    patch.set_weights(base.get_weights())78    return patch79AlignedXception71 = partial(80    wrap_bone,81    aligned_xception.Xception71,82    aligned_xception.preprocess_input)83def AlignedXception71Stride16(*args, **kwargs):84    base = AlignedXception71(*args, **kwargs)85    conf = base.get_config()86    conf = patch_config(conf, ['exit_flow/block1/unit1/sepconv3_depthwise'], 'strides', (1, 1))87    conf = patch_config(conf, ['exit_flow/block1/unit1/shortcut'], 'strides', (1, 1))88    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv1_depthwise'], 'dilation_rate', (2, 2))89    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv2_depthwise'], 'dilation_rate', (2, 2))90    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv3_depthwise'], 'dilation_rate', (2, 2))91    patch = models.Model.from_config(conf)92    patch.set_weights(base.get_weights())93    return patch94def AlignedXception71Stride8(*args, **kwargs):95    base = AlignedXception71(*args, **kwargs)96    conf = base.get_config()97    conf = patch_config(conf, ['entry_flow/block5/unit1/sepconv3_depthwise'], 'strides', (1, 1))98    conf = patch_config(conf, ['entry_flow/block5/unit1/shortcut'], 'strides', (1, 1))99    for i in range(16):100        conf = patch_config(conf, [101            'middle_flow/block1/unit{}/sepconv1_depthwise'.format(i + 1)], 'dilation_rate', (2, 2))102        conf = patch_config(conf, [103            'middle_flow/block1/unit{}/sepconv2_depthwise'.format(i + 1)], 'dilation_rate', (2, 2))104        conf = patch_config(conf, [105            'middle_flow/block1/unit{}/sepconv3_depthwise'.format(i + 1)], 'dilation_rate', (2, 2))106    conf = patch_config(conf, ['exit_flow/block1/unit1/sepconv3_depthwise'], 'strides', (1, 1))107    conf = patch_config(conf, ['exit_flow/block1/unit1/shortcut'], 'strides', (1, 1))108    # Output shape is different when using exit dilation rates like in109    # https://github.com/bonlime/keras-deeplab-v3-plus/blob/master/model.py#L265110    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv1_depthwise'], 'dilation_rate', (2, 2))111    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv2_depthwise'], 'dilation_rate', (2, 2))112    conf = patch_config(conf, ['exit_flow/block2/unit1/sepconv3_depthwise'], 'dilation_rate', (2, 2))113    patch = models.Model.from_config(conf)114    patch.set_weights(base.get_weights())...

luna_s3_p8.py

Source:luna_s3_p8.py

1import data_transforms2import data_iterators3import pathfinder4import utils5import string6import numpy as np7import lasagne as nn8# IMPORT A CORRECT PATCH MODEL HERE9import lung_segmentation10import configs_seg_patch.luna_p8 as patch_config11# print utils.get_script_name(__file__).split('_')[-1]12# print patch_config.__name__.split('.')[-1]13# assert utils.get_script_name(__file__).replace('_s*_', '_') == patch_config.__name__.split('.')[-1]14rng = patch_config.rng15# calculate the following things correctly!16p_transform = {'patch_size': (416, 416, 416),17               'mm_patch_size': (416, 416, 416),18               'pixel_spacing': patch_config.p_transform['pixel_spacing']19               }20window_size = 16021stride = 12822n_windows = (p_transform['patch_size'][0] - window_size) / stride + 123valid_pids = patch_config.valid_pids24def data_prep_function(data, luna_annotations, pixel_spacing, luna_origin,25                       p_transform=p_transform,26                       p_transform_augment=None):27    # make sure the data is processed the same way28    lung_mask = lung_segmentation.segment_HU_scan(data)29    x, annotations_tf, tf_matrix, lung_mask_out = data_transforms.transform_scan3d(data=data,30                                                                                   pixel_spacing=pixel_spacing,31                                                                                   p_transform=p_transform,32                                                                                   luna_annotations=luna_annotations,33                                                                                   p_transform_augment=None,34                                                                                   luna_origin=luna_origin,35                                                                                   lung_mask=lung_mask)36    x = data_transforms.pixelnormHU(x)37    y = data_transforms.make_3d_mask_from_annotations(img_shape=x.shape, annotations=annotations_tf, shape='sphere')38    return x, y, lung_mask_out, annotations_tf, tf_matrix39valid_data_iterator = data_iterators.LunaScanPositiveLungMaskDataGenerator(data_path=pathfinder.LUNA_DATA_PATH,40                                                                           transform_params=p_transform,41                                                                           data_prep_fun=data_prep_function,42                                                                           rng=rng,43                                                                           batch_size=1,44                                                                           patient_ids=valid_pids,45                                                                           full_batch=True,46                                                                           random=False, infinite=False)47def build_model():48    metadata_dir = utils.get_dir_path('models', pathfinder.METADATA_PATH)49    metadata_path = utils.find_model_metadata(metadata_dir, patch_config.__name__.split('.')[-1])50    metadata = utils.load_pkl(metadata_path)51    print 'Build model'52    model = patch_config.build_model(patch_size=(window_size, window_size, window_size))53    all_layers = nn.layers.get_all_layers(model.l_out)54    num_params = nn.layers.count_params(model.l_out)55    print '  number of parameters: %d' % num_params56    print string.ljust('  layer output shapes:', 36),57    print string.ljust('#params:', 10),58    print 'output shape:'59    for layer in all_layers:60        name = string.ljust(layer.__class__.__name__, 32)61        num_param = sum([np.prod(p.get_value().shape) for p in layer.get_params()])62        num_param = string.ljust(num_param.__str__(), 10)63        print '    %s %s %s' % (name, num_param, layer.output_shape)64    nn.layers.set_all_param_values(model.l_out, metadata['param_values'])...

luna_s2_p8.py

Source:luna_s2_p8.py

1import data_transforms2import data_iterators3import pathfinder4import utils5import string6import numpy as np7import lasagne as nn8# IMPORT A CORRECT PATCH MODEL HERE9import configs_seg_patch.luna_p8 as patch_config10# print utils.get_script_name(__file__).split('_')[-1]11# print patch_config.__name__.split('.')[-1]12# assert utils.get_script_name(__file__).replace('_s*_', '_') == patch_config.__name__.split('.')[-1]13rng = patch_config.rng14# calculate the following things correctly!15p_transform = {'patch_size': (416, 416, 416),16               'mm_patch_size': (416, 416, 416),17               'pixel_spacing': patch_config.p_transform['pixel_spacing']18               }19window_size = 16020stride = 12821n_windows = (p_transform['patch_size'][0] - window_size) / stride + 122valid_pids = patch_config.valid_pids23def data_prep_function(data, luna_annotations, pixel_spacing, luna_origin,24                       p_transform=p_transform,25                       p_transform_augment=None):26    # make sure the data is processed the same way 27    x, annotations_tf, tf_matrix = data_transforms.transform_scan3d(data=data,28                                                                    pixel_spacing=pixel_spacing,29                                                                    p_transform=p_transform,30                                                                    luna_annotations=luna_annotations,31                                                                    p_transform_augment=None,32                                                                    luna_origin=luna_origin)33    x = data_transforms.pixelnormHU(x)34    y = data_transforms.make_3d_mask_from_annotations(img_shape=x.shape, annotations=annotations_tf, shape='sphere')35    return x, y, annotations_tf, tf_matrix36valid_data_iterator = data_iterators.LunaScanPositiveDataGenerator(data_path=pathfinder.LUNA_DATA_PATH,37                                                                   transform_params=p_transform,38                                                                   data_prep_fun=data_prep_function,39                                                                   rng=rng,40                                                                   patient_ids=valid_pids,41                                                                   random=False, infinite=False)42def build_model():43    metadata_dir = utils.get_dir_path('models', pathfinder.METADATA_PATH)44    metadata_path = utils.find_model_metadata(metadata_dir, patch_config.__name__.split('.')[-1])45    metadata = utils.load_pkl(metadata_path)46    print 'Build model'47    model = patch_config.build_model(patch_size=(window_size, window_size, window_size))48    all_layers = nn.layers.get_all_layers(model.l_out)49    num_params = nn.layers.count_params(model.l_out)50    print '  number of parameters: %d' % num_params51    print string.ljust('  layer output shapes:', 36),52    print string.ljust('#params:', 10),53    print 'output shape:'54    for layer in all_layers:55        name = string.ljust(layer.__class__.__name__, 32)56        num_param = sum([np.prod(p.get_value().shape) for p in layer.get_params()])57        num_param = string.ljust(num_param.__str__(), 10)58        print '    %s %s %s' % (name, num_param, layer.output_shape)59    nn.layers.set_all_param_values(model.l_out, metadata['param_values'])...

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.