Skip to content
This repository has been archived by the owner on Dec 1, 2021. It is now read-only.

In case of really small dataset. MAX_STEPS can be less than 1 and lead to error. #1194

Closed
oatawa1 opened this issue Sep 2, 2020 · 0 comments
Closed

In case of really small dataset. MAX_STEPS can be less than 1 and lead to error. #1194

oatawa1 opened this issue Sep 2, 2020 · 0 comments
Assignees
@oatawa1
Labels
bug Something isn't working

Comments

@oatawa1
Copy link
Contributor

oatawa1 commented Sep 2, 2020 •
edited
Loading

From this line of code

blueoil/blueoil/cmd/train.py

Line 194 in ad162bd

max_steps = int(train_dataset.num_per_epoch / config.BATCH_SIZE * config.MAX_EPOCHS)

When the dataset is really small, precisely, no. of train dataset*max_epochs < batch_size, it can cause max_steps = 0 which leads to OverflowError.

  File "blueoil/cmd/train.py", line 200, in start_training
    progbar.update(last_step)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/utils/generic_utils.py", line 396, in update
    numdigits = int(np.log10(self.target)) + 1
OverflowError: cannot convert float infinity to integer
@oatawa1 oatawa1 added the bug Something isn't working label Sep 2, 2020
@oatawa1 oatawa1 self-assigned this Sep 2, 2020
@oatawa1 oatawa1 mentioned this issue Sep 2, 2020
Merged
bo-mergebot bot pushed a commit that referenced this issue Sep 3, 2020
@oatawa1
Fix MAX_STEPS = 0 error by add suggestion and exit (#1195)
2b80ed1 
## What this patch does to fix the issue.
Change the minimum no. of `MAX_STEPS` to 1.

## Link to any relevant issues or pull requests.
#1194
@oatawa1 oatawa1 closed this as completed Sep 4, 2020
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees

@oatawa1 oatawa1

Labels
bug Something isn't working
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@oatawa1