uploader: validate tensors before uploading them #3624
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ericdnielsen
approved these changes
May 12, 2020
caisq
added a commit
to caisq/tensorboard
that referenced
this pull request
May 19, 2020
* Motivation for features / changes
* Guard against malformed tensor data points in the uploader: fail during
upload time instead of serving time.
* Technical description of changes
* In `_TensorBatchedRequestSender`, try to parse a `TensorProto` before sending it
to tensorboard.dev server.
* Add unit tests for the catching of malformed tensors.
* Some existing unit tests are modified. They previously used malformed `TensorProto` values.
* Screenshots of UI changes
* 
* Detailed steps to verify changes work correctly (as executed by you)
* Unit tests added, as stated above.
* Verified error catching using a real logdir with malformed histogram tensors.
* Performance benchmarks:
* I ran the uploader 8 times, using the stock histograms_demo logdir.
* 4 times with and 4 times without tensor validation, in interleaved order.
* Total time it takes to finish uploading the logdir:
* With validation: 80.3±0.03 s (mean±SD)
* Without validation: 80.2±0.004 s
* Absolute diff: 0.033 s
* Relative diff: 0.04%
* Independent samples t-test: t = 1.88, p = 0.108
caisq
added a commit
that referenced
this pull request
May 27, 2020
* Motivation for features / changes
* Guard against malformed tensor data points in the uploader: fail during
upload time instead of serving time.
* Technical description of changes
* In `_TensorBatchedRequestSender`, try to parse a `TensorProto` before sending it
to tensorboard.dev server.
* Add unit tests for the catching of malformed tensors.
* Some existing unit tests are modified. They previously used malformed `TensorProto` values.
* Screenshots of UI changes
* 
* Detailed steps to verify changes work correctly (as executed by you)
* Unit tests added, as stated above.
* Verified error catching using a real logdir with malformed histogram tensors.
* Performance benchmarks:
* I ran the uploader 8 times, using the stock histograms_demo logdir.
* 4 times with and 4 times without tensor validation, in interleaved order.
* Total time it takes to finish uploading the logdir:
* With validation: 80.3±0.03 s (mean±SD)
* Without validation: 80.2±0.004 s
* Absolute diff: 0.033 s
* Relative diff: 0.04%
* Independent samples t-test: t = 1.88, p = 0.108
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
upload time instead of serving time.
_TensorBatchedRequestSender, try to parse aTensorProtobefore sending itto tensorboard.dev server.
TensorProtovalues.