-
Notifications
You must be signed in to change notification settings - Fork 78
Home
Yuxuan Liu edited this page Feb 1, 2021
·
7 revisions
They are functions that conduct evaluation or complete a step of training.
The folder for these functions
- trainers: trainers can accept different types of data (directly from datasets' collate_fn) based on the implementations. But other inputs are basically fixed.
Please refer to function documents of compound_annotation visualDet3D/utils/utils.py. This is a legacy function from RetinaNet.
Link for the function inside the current master version
You could first have a look at the heads inside this repo
Suggestions:
- Write a new subclass; modify the init_layers method for new structure.
- Write a completely new Head class, keeping forward, get_bboxes and loss methods as main interfaces and modify all other contents inside. (If you need to remove anchors, you may need to write a new Detector class. Just like mmdetection)
Main design ideas for data augmentations in this repo:
- For training 3D detection, we should also preprocess calibration matrix
$P$ when we are preprocessing the image/labels. These are all done with in augmentation pipeline. - Models are trained and completing inferences with augmented image and calibration matrixes.
- 2D-related prediction results are corrected by aligning the augmented
$P$ to the original$P$ in evaluator pipelines
There could be better way to deal with this problem, but the above are for the current repo.