-
Notifications
You must be signed in to change notification settings - Fork 340
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Auto Micro Batch] After enable Auto Micro Batch feature in WDL of modelzoo, but get an error. #127
Comments
This was referenced Apr 1, 2022
Closed
This was referenced Apr 11, 2022
This was referenced May 13, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I want to enable Auto Micro Batch feature in WDL and follow the steps in DeepRec Docs, but I get an error.
Code to reproduce the issue
I use following codes to enable Auto Graph Fusion. The full code please see Full code
Run
python train.py --steps 1000 --no_eval --micro_batch 2
can reproduce error. Use WDL dataset.When set
--micro_batch
(micro_batch_num) to 1, it's OK."AutoMicroBatch功能依赖于用户开启图优化的选项" means
Auto Graph Fusion
? It can be enabled by--op_fusion True
, but get the same error. And I also get terrible in enablingAuto Graph Fusion
, see issue #126This seems to be because of the initialization of dataset in MonitorTrainingSession. So this issue is different from #86 which use tf.Session().
logs
The text was updated successfully, but these errors were encountered: