[Bug]: setting max_epochs when engine instantiation not works

[shanmugamani]

Describe the bug

i tried to create model using custom dataset ,so ,
i used following code

Import the datamodule

from anomalib.data import Folder
from anomalib.data.utils import TestSplitMode

Create the datamodule

datamodule = Folder(
name="bottle",
root="datasets/bottle",
normal_dir="good",
test_split_mode=TestSplitMode.SYNTHETIC,
task="classification",
image_size=(512,512)
)

Setup the datamodule

datamodule.setup(),

after that # Import the model and engine
from anomalib.models import Patchcore
from anomalib.engine import Engine

Create the model and engine

model = Patchcore()
engine = Engine(task="classification",
accelerator="auto",
check_val_every_n_epoch=1,
max_epochs=2,
num_sanity_val_steps=0,
val_check_interval=1.0)
,i configured train arguments like this,
i gave max_epochs=2,
but it only run 1 time ?? how to resolve this issue ?

Dataset

Other (please specify in the text field below)

Model

PatchCore

Steps to reproduce the behavior

i tried to create model using custom dataset ,so ,
i used following code

Import the datamodule

from anomalib.data import Folder
from anomalib.data.utils import TestSplitMode

Create the datamodule

datamodule = Folder(
name="bottle",
root="datasets/bottle",
normal_dir="good",
test_split_mode=TestSplitMode.SYNTHETIC,
task="classification",
image_size=(512,512)
)

Setup the datamodule

datamodule.setup(),

after that # Import the model and engine
from anomalib.models import Patchcore
from anomalib.engine import Engine

Create the model and engine

model = Patchcore()
engine = Engine(task="classification",
accelerator="auto",
check_val_every_n_epoch=1,
max_epochs=2,
num_sanity_val_steps=0,
val_check_interval=1.0)
,i configured train arguments like this,
i gave max_epochs=2,
but it only run 1 time ?? how to resolve this issue ?

OS information

OS information:

• OS: [e.g. Ubuntu 20.04]
• Python version: [e.g. 3.10.0]
• Anomalib version: [e.g. 0.3.6]
• PyTorch version: [e.g. 1.9.0]
• CUDA/cuDNN version: [e.g. 11.1]
• GPU models and configuration: [e.g. 2x GeForce RTX 3090]
• Any other relevant information: [e.g. I'm using a custom dataset]

Expected behavior

need to run 2 times

Screenshots

Pip/GitHub

pip

What version/branch did you use?

No response

Configuration YAML

null

Logs

null

Code of Conduct

• I agree to follow this project's Code of Conduct

[Samet Akcay]

@shanmugamani1023, Patchcore only needs one epoch to extract the features. Setting more epochs does not help training, which is the reason it is 1 by default and cannot be changed.

For more details, you could refer to the paper
https://arxiv.org/abs/2106.08265

I'm closing this since this is not an issue. Thanks

[shanmugamani]

Thanks, so does every model have its own predefined epochs? Therefore,
Could you please provide me with the epoch values for each model?
Which model allows me to change the epoch number?

[Samet Akcay]

You could check them from either model config

anomalib/configs/model/efficient_ad.yaml

Line 17 in 7a963e9

 max_epochs: 200 

or the model implementation

anomalib/src/anomalib/models/image/patchcore/lightning_model.py

Line 118 in 7a963e9

 return {"gradient_clip_val": 0, "max_epochs": 1, "num_sanity_val_steps": 0} 

[shanmugamani]

Thank you so much, its works fine.