This repository has been archived by the owner on Dec 7, 2021. It is now read-only.
fix: test asset distribution to include all tags on test/train split #823
Conversation
This file contains 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
|
@hermanho please re-target your PR to the |
The test asset may not included all tags when export with test/train split option in current venison (2.1.0).
hermanho
force-pushed
the
fix-test-train-split-sampling
branch
from
52f528a to
c609dc6
Compare
Codecov Report
@@ Coverage Diff @@
## master #823 +/- ##
==========================================
- Coverage 87.9% 87.76% -0.14%
==========================================
Files 129 129
Lines 4845 4880 +35
Branches 922 929 +7
==========================================
+ Hits 4259 4283 +24
- Misses 584 595 +11
Partials 2 2
Continue to review full report at Codecov.
|
|
@tbarlow12 changed |
wbreza
approved these changes
Aug 22, 2019
There was a problem hiding this comment.
Nice job on this - especially the shared logic used between Pascal VOC & CNTK. @JacopoMangiavacchi Can you please take a pass on these changes.
src/providers/export/cntk.ts
Outdated
| @@ -3,6 +3,7 @@ import { ExportProvider, IExportResults } from "./exportProvider"; | |||
| import { IAssetMetadata, IExportProviderOptions, IProject } from "../../models/applicationState"; | |||
| import HtmlFileReader from "../../common/htmlFileReader"; | |||
| import Guard from "../../common/guard"; | |||
| import {splitTestAsset} from "./testAssetsSplitHelper"; | |||
There was a problem hiding this comment.
Nit: Spacing around { in import statements.
| export function splitTestAsset(allAssets: IAssetMetadata[], tags: ITag[], testSplitRatio: number): string[] { | ||
| const testAssets: string[] = []; | ||
|
|
||
| if (testSplitRatio > 0 && testSplitRatio <= 1) { |
There was a problem hiding this comment.
Consider inverting if statement to reduce code nesting.
JacopoMangiavacchi
approved these changes
Aug 23, 2019
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.
The test asset may not included all tags when export with test/train split option and uneven distribution of tags in current venison (2.1.0). Here is an example:
Image 1-8 are tagged with Tag-1.
Image 9-10 are tagged with Tag-2.
Export provider: PascalVOC
Asset State: Only tagged Assets
Test / Train Split is 80%.
Output
tag-1_val.txt
tag-2_val.txt
In current venison, there is no images of tag "Tag-1" in the text file "tag-1_val.txt" because the split method is based on the whole image set so only image 9 & 10 are put in test asset. The split method is changed to extract the test assets by each tag and combine them at the end. Image 7,8,10 will put in test asset after the change applied.
Output
tag-1_val.txt
tag-2_val.txt