about continue training

[zrt791521360]

Hello, thank you for your excellent work：
Because the computer is unstable, there will be interruptions, so i need to continue training.
I was confused by the many save parameters in the option file，In order to describe my problem clearly, please forgive me for using Chinese.
继续训练是不是加上--continue_train,再修改--epoch_count为上次运行到的epoch就可以了？我看还有其他iter保存参数什么的，被弄晕了

[zrt791521360]

Just now I made a mistake about the learning rate when i try to continue training

[zrt791521360]

could you please give me some details about continuing training? Thank you!

[JunlinHan]

Hi,
You are correct! Just add --continue_train --epoch_count XX( where XX is the epoch in your automatically saved latest model)