https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.3/docs/MODEL_ZOO_cn.md
训练策略
- 我们采用和Detectron相同的训练策略。
- 1x 策略表示：在总batch size为8时，初始学习率为0.01，在8 epoch和11 epoch后学习率分别下降10倍，最终训练12 epoch。
- 2x 策略为1x策略的两倍，同时学习率调整位置也为1x的两倍。
Training Schedules
We use three training schedules, indicated by the lr schd column in the tables below.
-
1x: For minibatch size 16, this schedule starts at a LR of 0.02 and is decreased by a factor of * 0.1 after 60k and 80k iterations and finally terminates at 90k iterations. This schedules results in 12.17 epochs over the 118,287 images in
coco_2014_trainunioncoco_2014_valminusminival(or equivalently,coco_2017_train). - 2x: Twice as long as the 1x schedule with the LR change points scaled proportionally.
- s1x ("stretched 1x"): This schedule scales the 1x schedule by roughly 1.44x, but also extends the duration of the first learning rate. With a minibatch size of 16, it reduces the LR by * 0.1 at 100k and 120k iterations, finally ending after 130k iterations.
All training schedules also use a 500 iteration linear learning rate warm up. When changing the minibatch size between 8 and 16 images, we adjust the number of SGD iterations and the base learning rate according to the principles outlined in our paper Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour.
