aws / aws/sagemaker-tensorflow-extensions
PipeModeDataset and tf.Dataset.shard
Open
- Dominant language
- C++
- Stars
- 54
- Forks
- 39
- PR merge metrics
- No merged PRs in 30d
Description
I have gotten errors trying to shard my dataset over multiple gpu's using PipeMode. OutOfRange error usually comes up on one or more of them. Does anyone have an example of this working?
Contributor guide
Assessment
This issue has not been assessed yet.