-
Notifications
You must be signed in to change notification settings - Fork 74k
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
Using Dataset api with Estimator in MirroredStrategy, Non-DMA-safe string tensor error #19588
Comments
I meet the same problem @skye in using object detection apis |
Can you provide code to repro the problem? |
Rohan/Priya: I'm guessing this is what happens when a |
def get_inputs(mode, csv_file, batch_size, label_list, preprocess):
distribution = tf.contrib.distribute.MirroredStrategy() classifier = tf.estimator.Estimator( |
Nothing special, its just common code with MirroredStrategy. |
Nagging Assignees @rohan100jain, @guptapriya: It has been 14 days with no activity and this issue has an assignee. Please update the label and/or status accordingly. |
As Derek mentioned, currently we don't have a mechanism for specifying if some outputs should be in host memory and we assume (to a large extent) that they'd be in device memory. Strings can't be in device memory, hence the bug. I shall work on having a dynamic method of identifying which outputs should be allocated on the host / device. Stay tuned for a fix in a bit. |
FunctionBufferingResource. This allows for types such as strings that are always in host memory to be returned from the FunctionBufferingResource. Fixes tensorflow#19588 PiperOrigin-RevId: 202206052
@rohan100jain can this issue be closed now that your fix has been merged? |
System information
Describe the problem
Using mutilple gpu by MirroredStrategy, Get ' Non-DMA-safe string tensor may not be copied from/to a GPU.' error
Source code / logs
The text was updated successfully, but these errors were encountered: