I am using TF-slim to train my own dataset. According to the tutorial, in the finetune_inception_v1_on_flowers.sh, I set the checkpoint path as ./inception_resnet_v2.ckpt, and get lots of new checkpoints in another file folder, and the latest files are model.ckpt-332331.data-00000-of-00001,model.ckpt-332331.index, model.ckpt-332331.meta.
now I want to change the learning rate to fine turing the latest check point, but I don't know how to load them in the finetune_inception_v1_on_flowers.sh because there are 3 files instead of one.
firstly, I tried to copy the model.ckpt-332331.data-00000-of-00001 and set the checkpoint path as ./model.ckpt-332331.data-00000-of-00001 ,but there is the error:
NotFoundError (see above for traceback): Key InceptionResnetV2/AuxLogits/Conv2d_1b_1x1/weights/Adam not found in checkpoint [[Node: save/RestoreV2_4 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_recv_save/Const_0, save/RestoreV2_4/tensor_names, save/RestoreV2_4/shape_and_slices)]]
secondly, I removed the files from log directory ,then the error is :
DataLossError (see above for traceback): Unable to open table file /media/re/7A308F7E308F405D/xsj/checkpoints/model.ckpt-332331.data-00000-of-00001: Data loss: not an sstable (bad magic number): perhaps your file is in a different file format and you need to use a different restore operator? [[Node: save/RestoreV2_5 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_recv_save/Const_0, save/RestoreV2_5/tensor_names, save/RestoreV2_5/shape_and_slices)]]
Can anyone help me to solve the problem? Thank you very much!
As mentioned in https://www.tensorflow.org/guide/saved_model
There is not a physical file called /tmp/model.ckpt. It is the prefix of filenames created for the checkpoint. Users only interact with the prefix instead of physical checkpoint files.
So you just need to set checkpoint path as ./model.ckpt-332331
进行微调时,只需指定目录($ CHECKPOINT_PATH)而不是任何特定文件,它将自动获取路径。
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.