tensorflow / tensorflow/models

Unable to create/update the checkpoints during training

Open
#9,168 1 comment 0 reactions 3 assignees View on GitHub

@pkulzc is already working on this.

Since Aug 31, 2020.

models:research:odapi type:bug
Dominant language
Python
Stars
77.7k
Forks
44.8k
PR merge metrics
No merged PRs in 30d

Description

Hi team,

1. The entire URL of the file you are using

https://github.com/tensorflow/models/blob/master/research/object_detection/configs/tf2/faster_rcnn_resnet101_v1_640x640_coco17_tpu-8.config

2. Describe the bug

I'm using the faster RCNN 640x640 model for object detection on a custom dataset. During the training, I see that the events file, events.out.tfevents.1598792593.Meghana.23775.1504.v2 inside thetrain directory is getting updated. But, the corresponding checkpoints are not created.

Refer to the screenshots for the timestamps

Screenshot from 2020-08-30 20-14-40
image

The checkpoints in the above screenshots refer to the restored checkpoints. Not the once created during the training process.
Also, I see no log messages stating that the checkpoints are being saved as in TF1
image

Versions:
OS: ubuntu
TF: 2.3.0
python: 3.8.3

Thank you,
Meghana

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.