beenull/ente
1
0
Fork 0
Code Issues Pull requests Actions Packages Projects Releases Wiki Activity

[mob][photos] Big cluster suggestions first

Browse source
This commit is contained in:
laurenspriem 2024-05-11 15:20:28 +05:30
parent abff589c21
commit fc6830bdaf
1 changed files with 5 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
mobile/lib/services/machine_learning/face_ml/feedback/cluster_feedback.dart
View file

@ -335,7 +335,6 @@ class ClusterFeedbackService {
for (final clusterID in clustersToInspect) {
final int originalClusterSize = allClusterToFaceCount[clusterID]!;
final faceIDs = await faceMlDb.getFaceIDsForCluster(clusterID);
final originalFaceIDsSet = faceIDs.toSet();
final embeddings = await faceMlDb.getFaceEmbeddingMapForFaces(faceIDs);
@ -645,7 +644,7 @@ class ClusterFeedbackService {
double maxMedianDistance = 0.62,
double goodMedianDistance = 0.55,
double maxMeanDistance = 0.65,
double goodMeanDistance = 0.40,
double goodMeanDistance = 0.45,
}) async {
final w = (kDebugMode ? EnteWatch('getSuggestions') : null)?..start();
// Get all the cluster data
@ -667,11 +666,12 @@ class ClusterFeedbackService {
final smallestPersonClusterSize = personClusters
.map((clusterID) => allClusterIdsToCountMap[clusterID] ?? 0)
.reduce((value, element) => min(value, element));
final checkSizes = [20, kMinimumClusterSizeSearchResult, 10, 5, 1];
final checkSizes = [100, 20, kMinimumClusterSizeSearchResult, 10, 5, 1];
late Map<int, Vector> clusterAvgBigClusters;
final List<(int, double)> suggestionsMean = [];
for (final minimumSize in checkSizes.toSet()) {
if (smallestPersonClusterSize >= minimumSize) {
if (smallestPersonClusterSize >=
min(minimumSize, kMinimumClusterSizeSearchResult)) {
clusterAvgBigClusters = await _getUpdateClusterAvg(
allClusterIdsToCountMap,
ignoredClusters,
@ -685,7 +685,7 @@ class ClusterFeedbackService {
clusterAvgBigClusters,
personClusters,
ignoredClusters,
goodMeanDistance,
(minimumSize == 100) ? goodMeanDistance + 0.05 : goodMeanDistance,
);
w?.log(
'Calculate suggestions using mean for ${clusterAvgBigClusters.length} clusters of min size $minimumSize',