0ct0pu5/ente
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

[mob][photos] Correct suggestion logic again

Browse source
This commit is contained in:
laurenspriem 2024-04-24 17:06:35 +05:30
parent 462d1d4854
commit e0fbb2620b
1 changed files with 7 additions and 28 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

35
mobile/lib/services/machine_learning/face_ml/feedback/cluster_feedback.dart
View file

@ -455,7 +455,7 @@ class ClusterFeedbackService {
double maxMedianDistance = 0.62,
double goodMedianDistance = 0.55,
double maxMeanDistance = 0.65,
double goodMeanDistance = 0.54,
double goodMeanDistance = 0.50,
}) async {
final w = (kDebugMode ? EnteWatch('getSuggestions') : null)?..start();
// Get all the cluster data
@ -472,10 +472,10 @@ class ClusterFeedbackService {
.map((clusterID) => allClusterIdsToCountMap[clusterID] ?? 0)
.reduce((value, element) => min(value, element));
final checkSizes = [kMinimumClusterSizeSearchResult, 20, 10, 5, 1];
late final Map<int, Vector> clusterAvgBigClusters;
for (final minimumSize in checkSizes.toSet()) {
if (smallestPersonClusterSize >= minimumSize) {
final Map<int, Vector> clusterAvgBigClusters =
await _getUpdateClusterAvg(
clusterAvgBigClusters = await _getUpdateClusterAvg(
allClusterIdsToCountMap,
ignoredClusters,
minClusterSize: minimumSize,
@ -502,29 +502,8 @@ class ClusterFeedbackService {
}
w?.reset();
// Get and update the cluster summary to get the avg (centroid) and count
final Map<int, Vector> clusterAvg = await _getUpdateClusterAvg(
allClusterIdsToCountMap,
ignoredClusters,
);
w?.log(
'computed avg for ${clusterAvg.length} clusters,',
);
// Find the other cluster candidates based on the mean
final List<(int, double)> suggestionsMean = _calcSuggestionsMean(
clusterAvg,
personClusters,
ignoredClusters,
goodMeanDistance,
);
if (suggestionsMean.isNotEmpty) {
return suggestionsMean
.map((e) => (e.$1, e.$2, true))
.toList(growable: false);
}
// Find the other cluster candidates based on the median
final clusterAvg = clusterAvgBigClusters;
final List<(int, double)> moreSuggestionsMean = _calcSuggestionsMean(
clusterAvg,
personClusters,
@ -595,7 +574,7 @@ class ClusterFeedbackService {
final List<double> distances = [];
for (final otherEmbedding in sampledOtherEmbeddings) {
for (final embedding in sampledEmbeddings) {
distances.add(cosineDistanceSIMD(embedding,otherEmbedding));
distances.add(cosineDistanceSIMD(embedding, otherEmbedding));
}
}
distances.sort();
@ -800,7 +779,7 @@ class ClusterFeedbackService {
continue;
}
final Vector avg = clusterAvg[personCluster]!;
final distance = cosineDistanceSIMD(avg,otherAvg);
final distance = cosineDistanceSIMD(avg, otherAvg);
if (distance < maxClusterDistance) {
if (minDistance == null || distance < minDistance) {
minDistance = distance;
@ -951,7 +930,7 @@ class ClusterFeedbackService {
final fileIdToDistanceMap = {};
for (final entry in faceIdToVectorMap.entries) {
fileIdToDistanceMap[getFileIdFromFaceId(entry.key)] =
cosineDistanceSIMD(personAvg,entry.value);
cosineDistanceSIMD(personAvg, entry.value);
}
w?.log('calculated distances for cluster $clusterID');
suggestion.filesInCluster.sort((b, a) {