Fzy-Sy

About resume training

5

if args.restore_ckpt is not None: strStep = os.path.split(args.restore_ckpt)[-1].split('_')[0] total_steps = int(strStep) if strStep.isdigit() else 0 else: total_steps = 0 Hi, I found this code in train.py. The class OneCycleLR need...

Using C+T for initialization, so, how many training steps and crop size are set when training RFAB in the stage of M-step, and which datasets are used to validate?

I remember that your work was submitted on the official websites of Sintel and KITTI, why can't I find it now?

thanks for your good work, I have some questions for M-step 1. Using C+T to generate new dataset G1. 2. Using G1 to train RAFT get new weight C+T+G1. 3....

请问M-step，RAFT的训练是在新数据集上只执行一个epoch还是迭代更多个epoch？RAFT原文中在MPI上是迭代100k次，M-step也是将RAFT在新数据集上迭代100K次吗？