mirror of https://github.com/hanzoai/anomalib.git synced 2026-10-30 18:48:11 +00:00

Files

T

History

The previous commit was meant to remove .github/workflows and removed
every tracked path instead — 892 files, the whole repo. This puts back
the 874 that were never in scope; the eighteen workflow files stay gone,
for the reasons that commit gives.

Content is byte-identical to 9d2aec9 outside .github/workflows.

2026-08-13 23:27:57 -07:00

gradio_inference.py

restore the tree

2026-08-13 23:27:57 -07:00

lightning_inference.py

restore the tree

2026-08-13 23:27:57 -07:00

openvino_inference.py

restore the tree

2026-08-13 23:27:57 -07:00

README.md

restore the tree

2026-08-13 23:27:57 -07:00

torch_inference.py

restore the tree

2026-08-13 23:27:57 -07:00

README.md

Inference

These files provide an example on how to use the trained (and optionally optimized) models.

Currently, you can use the use the torch, OpenVINO, and lightning models.

Lightning Inference

The lightning_inference.py provides an example on how to use the trained ckpt file to run inference on a dataset.

Example:

python tools/inference/lightning_inference.py \
  --model anomalib.models.Padim \
  --ckpt_path results/padim/mvtecad/bottle/weights/lightning/model.ckpt \
  --data.path datasets/MVTecAD/bottle/test/broken_large \
  --output ./outputs

You can also use a config file with the entrypoint

Here is a simple YAML file for Padim Model.

ckpt_path: results/padim/mvtecad/bottle/weights/lightning/model.ckpt
data:
  path: datasets/MVTecAD/bottle/test/broken_large
  transform: null
  image_size:
    - 256
    - 256
output: ./outputs
visualization_mode: simple
show: false
model:
  class_path: anomalib.models.Padim
  init_args:
    layers:
      - layer1
      - layer2
      - layer3
    input_size:
      - 256
      - 256
    backbone: resnet18
    pre_trained: true
    n_features: null

You can then use

python tools/inference/lightning_inference.py -c config.yaml