pyanomaly.core package

Subpackages

• pyanomaly.core.engine package
  • Subpackages
    • pyanomaly.core.engine.abstract package
      • Submodules
      • pyanomaly.core.engine.abstract.abstract_engine module
      • pyanomaly.core.engine.abstract.base_engine module
      • Module contents
    • pyanomaly.core.engine.functions package
      • Submodules
      • pyanomaly.core.engine.functions.amc module
      • pyanomaly.core.engine.functions.anopcn module
      • pyanomaly.core.engine.functions.anopred module
      • pyanomaly.core.engine.functions.memae module
      • pyanomaly.core.engine.functions.ocae module
      • pyanomaly.core.engine.functions.stae module
      • Module contents
  • Submodules
  • pyanomaly.core.engine.engine_api module
  • pyanomaly.core.engine.engine_registry module
  • pyanomaly.core.engine.utils module
  • Module contents
• pyanomaly.core.hook package
  • Subpackages
    • pyanomaly.core.hook.abstract package
      • Submodules
      • pyanomaly.core.hook.abstract.abstract_hook module
      • Module contents
    • pyanomaly.core.hook.functions package
      • Submodules
      • pyanomaly.core.hook.functions.amc_hooks module
      • pyanomaly.core.hook.functions.anopcn_hooks module
      • pyanomaly.core.hook.functions.anopred_hooks module
      • pyanomaly.core.hook.functions.base module
      • pyanomaly.core.hook.functions.memae_hooks module
      • pyanomaly.core.hook.functions.ocae_hooks module
      • pyanomaly.core.hook.functions.stae_hooks module
      • Module contents
  • Submodules
  • pyanomaly.core.hook.hook_registry module
  • pyanomaly.core.hook.hooks_api module
  • Module contents
• pyanomaly.core.optimizer package
  • Submodules
  • pyanomaly.core.optimizer.optimizer_api module
  • Module contents
• pyanomaly.core.other package
  • Submodules
  • pyanomaly.core.other.kmeans module
  • Module contents
• pyanomaly.core.scheduler package
  • Submodules
  • pyanomaly.core.scheduler.scheduler_api module
  • pyanomaly.core.scheduler.schedulers module
  • Module contents

Submodules

pyanomaly.core.utils module

@author: Yuhao Cheng @contact: yuhao.cheng[at]outlook.com

class pyanomaly.core.utils.AverageMeter(name='default')

Bases: object

Computes and store the average the current value

get_name()

update(val, n=1)

class pyanomaly.core.utils.ParamSet(name='default', **kwargs)

Bases: object

A set of a group of params with samiliar meaning

get_name()

get_params_names()

pyanomaly.core.utils.flow_batch_estimate(flow_model, tensor_batch, normalize, output_format='xym', optical_size=None, output_size=None)

output_format:

general: u,v xym: u,v,mag hsv: rgb:

pyanomaly.core.utils.frame_gradient(x)

The input is a video clip and get the gradient on the image Args:

x: the video with bs. [bs, d, 1, h, w]

Returns:

dx: [bs, d, 1, h, w] the gradient on x-axis dy: [bs, d, 1, h, w] the gradient on y-axis

pyanomaly.core.utils.get_batch_dets(det_model, batch_image)

Use the detecron2

pyanomaly.core.utils.grid_crop(bottom, bbox, object_size=(64, 64))

[ x2-x1 x1 + x2 - W + 1 ] [ ----- 0 --------------- ] [ W - 1 W - 1 ] [ ] [ y2-y1 y1 + y2 - H + 1 ] [ 0 ----- --------------- ] [ H - 1 H - 1 ]

pyanomaly.core.utils.image_gradient(image)

Args:

x: the image with bs. [bs,c,h,w]

Returns:

dx: [bs,c,h,w] the gradient on x-axis dy: [bs,c,h,w] the gradient on y-axis

pyanomaly.core.utils.make_info_message(current_step, max_step, model_type, batch_time, batch_size, data_time, loss_list)

pyanomaly.core.utils.modeldist(model)

To make the model in dist

pyanomaly.core.utils.modelparallel(model)

To make the model parallel

pyanomaly.core.utils.multi_obj_grid_crop(bottom, bbox, object_size=(64, 64))

[ x2-x1 x1 + x2 - W + 1 ] [ ----- 0 --------------- ] [ W - 1 W - 1 ] [ ] [ y2-y1 y1 + y2 - H + 1 ] [ 0 ----- --------------- ] [ H - 1 H - 1 ]

pyanomaly.core.utils.save_score_results(score, cfg, logger, verbose=None, config_name='None', current_step=0, time_stamp='time_step')

This method is used to store the normal/abnormal scores which are used for the evaluation functions The

pyanomaly.core.utils.tensorboard_vis_images(vis_objects, writer, global_step, normalize)

Visualize the images in tensorboard Args:

vis_objects: the dict of visualized images.{'name1':iamge, ...} writer: tensorboard global_step: the step normalize: {'use':..., 'mean':..., 'std':...}

pyanomaly.core.utils.tsne_vis(feature, feature_labels, vis_path)

pyanomaly.core.utils.vis_optical_flow(batch_optical, output_format, output_size, normalize)

Module contents

@author: Yuhao Cheng @contact: yuhao.cheng[at]outlook.com