Fix/anomaly scores normalization consistency (#156)

Feature/upgrade timm mlflow onnxconverter (#155)

build: Upgrade version to 2.7.3 (#154)

fix: Avoid unbound calls of mlflowclient properties (#152)

feat: Integrate mlflow tracking for sklearn based tasks (#151)

feat: Add custom_normalized_threshold parameter to AnomalibEvaluation (#150)

feat: upgrade smp and tensorboard packages (#148)

feat: Automate tag and release creation on merge to main (#147)

fix: add probabilities to sklearn classification test when gt is not available (#145)

Feature/torch 2.8.0 upgrade (#144)

fix: Fix inconsistent anomaly score (#141)

feat: Compress models uploaded to mlflow (#140)

feat: Torch 2.6.0 upgrade (#139)

build: Relax pydantic constraint (#138)

build: Allow installing pydantic >=1.10.10 (#137)