* Improved detect.py timing

* Eliminate 1 time_sync() call

* Inference-only time

* dash

* #Save section

* Cleanup

@KalenMike

* Fix `user_config_dir()` for GCP/AWS functions

Compatability fix for GCP functions and AWS lambda for user config directory in https://github.com/ultralytics/yolov5/pull/4628

* Windows skip check

* Changed onnx-simplifier check behavior

Export.py has been updated to check for onnx-simplifier requirement only when the --simplify argument is added.
Allows for better flexibility and one less requirement if simplify is not needed.

* Fix single-element tuples
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

* Allow specifying multi_label option for NMS when using torch hub

* Reformat
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

* adding get cropped functionality

* Add target logic in existing functions

* Crops cleanup

* Add dictionary keys: conf, cls, box

* Bug fixes - avoid return after first image
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

* Add `user_config_dir`

* Linux to .config

* Update `check_yaml()` comment

* Cleanup

Fix a bug when `file=''`

* Add suffix checks

* Cleanup

* Cleanup2

* Cleanup3

* added callbacks

* added back callback to main

* added save_dir to callback output

* reduced code count

* updated callbacks

* added default callback class to main, added missing parameters to on_model_save

* Glenn updates
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

* Fixed 'meta' and 'hyp' may out of order when using evolve

* Update gitignore

Fix 3 for Arial.ttf redownloads with hub inference, follow-on to #4628.

* Remove assert

* debug0

* trace=not opt.sync

* sync to sync_bn fix

* Cleanup

* Close plots

* Replace fig.close() for plt.close()

Layer 21 includes the information of xsmall objects

* Auto-download Arial.ttf on init

* Fix ROOT

* Add Annotator() class

* Download Arial

* 2x for loop

* Cleanup

* tuple 2 list

* max_size=1920

* bold logging results to

* tolist()

* im = annotator.im

* PIL save in detect.py

* Smart asarray in detect.py

* revert to cv2.imwrite

* Cleanup

* Return result asarray

* Add `Profile()` profiler

* CamelCase Timeout

* Resize after mosaic

* pillow>=8.0.0

* daemon imwrite

* Add cv2 support

* Remove plot_wh_methods and plot_one_box

* pil=False for hubconf.py annotations

* im.shape bug fix

* colorstr common.py

* join daemons

* Update t.daemon

* Removed daemon saving

* Add `Profile()` profiler

* CamelCase Timeout

* Initial commit

* Update

* Add EarlyStopping feature

* Add comment

* Cleanup

* Cleanup2

* debug

* debug2

* debug3

* debug3

* debug4

* debug5

* debug6

* debug7

* debug8

* debug9

* debug10

* debug11

* debug12

* Cleanup

* Add TODO for known DDP issue

This PR brings alignment in AP computation practices with Detectron2 and MMDetection. 

Problem first noted by @yusiyoh in https://github.com/ultralytics/yolov5/issues/4546

Auto-fix corrupt JPEGs PR introduced a bug whereby the f.seek() operation read all of the bytes in the image, resulting in the PIL image having nothing to read upon the .save() operation. 

Fix was to re-open the image using PIL before saving.