face01lib.damo_yolo.damo_internal.utils package

Submodules

face01lib.damo_yolo.damo_internal.utils.boxes module

adjust_box_anns(bbox, scale_ratio, padw, padh, w_max, h_max)

bboxes_iou(bboxes_a, bboxes_b, xyxy=True)

filter_box(output, scale_range)

output: (N, 5+class) shape

matrix_iou(a, b)

return iou of a and b, numpy version for data augenmentation

postprocess(cls_scores, bbox_preds, num_classes, conf_thre=0.7, nms_thre=0.45, imgs=None)

xyxy2cxcywh(bboxes)

xyxy2xywh(bboxes)

face01lib.damo_yolo.damo_internal.utils.checkpoint module

load_ckpt(model, ckpt)

save_checkpoint(state, is_best, save_dir, model_name='')

face01lib.damo_yolo.damo_internal.utils.debug_utils module

debug_input_vis(imgs, targets, ids, train_loader)

face01lib.damo_yolo.damo_internal.utils.demo_utils module

face01lib.damo_yolo.damo_internal.utils.dist module

This file contains primitives for multi-gpu communication. This is useful when doing distributed training.

all_gather(data, group=None)

Run all_gather on arbitrary picklable data (not necessarily tensors). :param data: any picklable object :param group: a torch process group. By default, will use a group which

contains all ranks on gloo backend.

戻り値:

list of data gathered from each rank

戻り値の型:

list[data]

gather(data, dst=0, group=None)

Run gather on arbitrary picklable data (not necessarily tensors). :param data: any picklable object :param dst: destination rank :type dst: int :param group: a torch process group. By default, will use a group which

contains all ranks on gloo backend.

戻り値:

on dst, a list of data gathered from each rank. Otherwise,

an empty list.

戻り値の型:

list[data]

get_local_rank() → int

戻り値:

The rank of the current process within the local (per-machine) process group.

get_local_size() → int

戻り値:

The size of the per-machine process group, i.e. the number of processes per machine.

get_num_devices()

get_rank() → int

get_world_size() → int

is_main_process() → bool

synchronize()

Helper function to synchronize (barrier) among all processes when using distributed training

time_synchronized()

pytorch-accurate time

wait_for_the_master(local_rank: int)

Make all processes waiting for the master to do some task.

face01lib.damo_yolo.damo_internal.utils.imports module

import_file(module_name, file_path, make_importable=False)

face01lib.damo_yolo.damo_internal.utils.logger module

class StreamToLoguru(level='INFO', caller_names=('apex', 'pycocotools'))

ベースクラス: object

stream object that redirects writes to a logger instance.

flush()

write(buf)

get_caller_name(depth=0)

パラメータ:

• depth (int) -- Depth of caller conext, use 0 for caller depth.

• value (Default) --

戻り値:

module name of the caller

戻り値の型:

str

redirect_sys_output(log_level='INFO')

setup_logger(save_dir, distributed_rank=0, mode='a')

setup logger for training and testing. :param save_dir: location to save log file :type save_dir: str :param distributed_rank: device rank when multi-gpu environment :type distributed_rank: int :param mode: log file write mode, append or override. default is a. :type mode: str

戻り値:

logger instance.

face01lib.damo_yolo.damo_internal.utils.metric module

class AverageMeter(window_size=50)

ベースクラス: object

Track a series of values and provide access to smoothed values over a window or the global series average.

property avg

clear()

property global_avg

property latest

property median

reset()

property total

update(value)

class MeterBuffer(window_size=20)

ベースクラス: defaultdict

Computes and stores the average and current value

clear_meters()

get_filtered_meter(filter_key='time')

reset()

update([E, ]**F) → None.  Update D from dict/iterable E and F.

If E is present and has a .keys() method, then does: for k in E: D[k] = E[k] If E is present and lacks a .keys() method, then does: for k, v in E: D[k] = v In either case, this is followed by: for k in F: D[k] = F[k]

get_total_and_free_memory_in_Mb(cuda_device)

gpu_mem_usage()

Compute the GPU memory usage for the current device (MB).

face01lib.damo_yolo.damo_internal.utils.model_utils module

fuse_conv_and_bn(conv, bn)

fuse_model(model)

get_model_info(model, tsize)

make_divisible(v, divisor=8, min_value=None)

replace_module(module, replaced_module_type, new_module_type, replace_func=None)

Replace given type in module to a new type. mostly used in deploy.

パラメータ:

• module (nn.Module) -- model to apply replace operation.

• replaced_module_type (Type) -- module type to be replaced.

• new_module_type (Type)

• replace_func (function) -- python function to describe replace logic. Defalut value None.

戻り値:

module that already been replaced.

戻り値の型:

model (nn.Module)

face01lib.damo_yolo.damo_internal.utils.timer module

class Timer

ベースクラス: object

add(time_diff)

property average_time

avg_time_str()

reset()

tic()

toc(average=True)

get_time_str(time_diff)

face01lib.damo_yolo.damo_internal.utils.visualize module

vis(img, boxes, scores, cls_ids, conf=0.5, class_names=None)

Module contents