This repository was archived by the owner on Jan 3, 2023. It is now read-only.
Mingshan/Adding resnet50 validation script #478
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
…naSystems/ngraph-tf into mingshan/validate_resnet50
…naSystems/ngraph-tf into mingshan/validate_resnet50
sayantan-nervana
suggested changes
Apr 17, 2019
| def check_validation_results(norm_dict, metric): | ||
| test_pass = True | ||
| for norm in norm_dict: | ||
| if norm_dict[norm] > 0.1: |
There was a problem hiding this comment.
so if we get ref accuracy = 75, and ng accuracy = 75.3, then is it a failure?
There was a problem hiding this comment.
This script is not comparing the accuracy. It compares the training loss value at every iteration
There was a problem hiding this comment.
Same thing for loss. If ref loss is 1, and we get 0.8, is the test passing?
| return total_loss, top1_acc, top5_acc | ||
|
|
||
|
|
||
| def parse_reference_file(filename): |
There was a problem hiding this comment.
parse_reference_file and parse_training_output can be a single function... I think they are separate because one parses a file, and the other parses string. Maybe we keep the string parsing function and just read the file into a string and reuse.
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
This PR added the validation script for resnet50 training with both synthetic data and real data.
The tf result references under tfGPU/ folder is collected running the same command in the script on TF GPU.
The patch to make the data loader for real data deterministic is also included, and also the patch to eliminate the average_loss encapsulates in the training graph.