feat: add native EfficientAD support - #165

Closed

DeepKnowledge1 wants to merge 23 commits into

mainfrom

feat/efficientad

Commits on Aug 27, 2026

feat: add EfficientAD algorithm package
DeepKnowledge1
committed
d6aa4bb
feat: implement EfficientAD with native AnomaVision interface
DeepKnowledge1
committed
8ca3122
fix: align EfficientAD student and teacher feature strides
DeepKnowledge1
committed
be5e7d1
fix: keep EfficientAD compatible with shared normalized dataset
DeepKnowledge1
committed
b42cc12
feat: expose EfficientAD from AnomaVision API
DeepKnowledge1
committed
99459b4
feat: route training through configurable EfficientAD implementation
DeepKnowledge1
committed
1f1b120
refactor: preserve training workflow while adding EfficientAD
DeepKnowledge1
committed
536ac36
feat: support model.name as a native algorithm selector
DeepKnowledge1
committed
0b0b8ec
docs: add EfficientAD configuration options and selector
DeepKnowledge1
committed
a29eeef
docs: add EfficientAD CPU example configuration
DeepKnowledge1
committed
f66626a
test: add EfficientAD fit and validation coverage
DeepKnowledge1
committed
c110401
docs: document native EfficientAD workflow
DeepKnowledge1
committed
fdfe45f
docs: clean EfficientAD documentation
DeepKnowledge1
committed
e915967
docs: document EfficientAD as a first-class algorithm
DeepKnowledge1
committed
d4e3ef8
test: validate native model.name algorithm selection
DeepKnowledge1
committed
fb2d93b
refactor: preserve existing training CLI and add EfficientAD routing
DeepKnowledge1
committed
d8923db
docs: add EfficientAD configuration reference
DeepKnowledge1
committed
73d365a
fix: make EfficientAD ONNX export free of data-dependent training check
DeepKnowledge1
committed
a2aad29
cofig
DeepKnowledge1
committed
c917a9e
fix(efficientad): make ONNX export skip data-dependent training check
DeepKnowledge1
committed
8945f40
fix: add EfficientAD inference threshold
DeepKnowledge1
committed
6ba2daf
docs: document EfficientAD-specific thresholding
DeepKnowledge1
committed
fd9908a
fix: lower EfficientAD detection threshold
DeepKnowledge1
committed
99ac8c8

Navigation Menu

feat: add native EfficientAD support - #165

feat: add native EfficientAD support#165
DeepKnowledge1 wants to merge 23 commits into
mainfrom
feat/efficientad

Commits on Aug 27, 2026

feat: add EfficientAD algorithm package

feat: implement EfficientAD with native AnomaVision interface

fix: align EfficientAD student and teacher feature strides

fix: keep EfficientAD compatible with shared normalized dataset

feat: expose EfficientAD from AnomaVision API

feat: route training through configurable EfficientAD implementation

refactor: preserve training workflow while adding EfficientAD

feat: support model.name as a native algorithm selector

docs: add EfficientAD configuration options and selector

docs: add EfficientAD CPU example configuration

test: add EfficientAD fit and validation coverage

docs: document native EfficientAD workflow

docs: clean EfficientAD documentation

docs: document EfficientAD as a first-class algorithm

test: validate native model.name algorithm selection

refactor: preserve existing training CLI and add EfficientAD routing

docs: add EfficientAD configuration reference

fix: make EfficientAD ONNX export free of data-dependent training check

cofig

fix(efficientad): make ONNX export skip data-dependent training check

fix: add EfficientAD inference threshold

docs: document EfficientAD-specific thresholding

fix: lower EfficientAD detection threshold