[英]Which checkpoint should I select for continue for Object detection training
I start to training until ckpt-7 then I stopped training.我开始训练直到 ckpt-7 然后我停止训练。 Then again I started training but befor I changed pipline config in fine tune chekpoint on my model.
然后我再次开始训练,但是在我更改了我的 model 上的微调检查点的管道配置之前。 I wrote latest check point and I changed its directory.
我写了最新的检查点,我改变了它的目录。 My loss function approximetly 0.899 before stopped to the training.
在停止训练之前,我的损失 function 大约为 0.899。 When I continue to train but its start to steps 100 and my loss fuction 15.009.
当我继续训练但它开始到步骤 100 和我的损失函数 15.009。
How can I contiune the model before stopped?在停止之前如何继续 model? What should I do?
我应该怎么办?
I am using centernet model with Colab.我在 Colab 中使用 centernet model。 Please explain I am new on that topic.
请解释我是该主题的新手。
I could understand your question that you could not resume the training where it stopped.我可以理解您的问题,即您无法在停止的地方恢复培训。
Actually with the updates in TF2, we need not change the finetune checkpoint parameter in the pipeline.config.实际上,随着 TF2 中的更新,我们无需更改 pipeline.config 中的微调检查点参数。 Re-run the same training script pointing to the same model_dir where your checkpoints are stored.
重新运行指向存储检查点的同一 model_dir 的相同训练脚本。
TF2 will automatically understand and resume from where the training stopped with the help of checkpoints created in the model_dir.在 model_dir 中创建的检查点的帮助下,TF2 将自动理解并从训练停止的位置恢复。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.