quadra 2.2.1__tar.gz → 2.2.3__tar.gz

{quadra-2.2.1 → quadra-2.2.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: quadra
-Version: 2.2.1
+Version: 2.2.3
 Summary: Deep Learning experiment orchestration library
 Home-page: https://orobix.github.io/quadra
 License: Apache-2.0
@@ -21,9 +21,9 @@ Classifier: Topic :: Software Development :: Libraries
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Provides-Extra: onnx
 Requires-Dist: albumentations (>=1.3,<1.4)
-Requires-Dist: anomalib-orobix (==0.7.0.dev140)
+Requires-Dist: anomalib-orobix (==0.7.0.dev141)
 Requires-Dist: boto3 (>=1.26,<1.27)
-Requires-Dist: grad_cam (==1.4.6)
+Requires-Dist: grad-cam-orobix (==1.5.3.dev001)
 Requires-Dist: h5py (>=3.8,<3.9)
 Requires-Dist: hydra_colorlog (>=1.2,<1.3)
 Requires-Dist: hydra_core (>=1.3,<1.4)
@@ -235,7 +235,7 @@ On the same dataset we can run a simple SSL training using the BYOL algorithm wi
 quadra experiment=generic/imagenette/ssl/byol logger=csv
 ```
 
-BYOL is not the only SSL algorithm available, you can find a list of all the available algorithms under `quadra/experiment/generic/imagenette/ssl` folder.
+BYOL is not the only SSL algorithm available, you can find a list of all the available algorithms under `quadra/configs/experiment/generic/imagenette/ssl` folder.
 
 
 ### Anomaly Detection Training
@@ -248,11 +248,11 @@ quadra experiment=generic/mnist/anomaly/padim logger=csv
 
 This will run an anomaly detection considering on of the classes as good (default is the number 9) and the rest as anomalies.
 
-This will make use of the [anomalib](https://github.com/openvinotoolkit/anomalib) library to train the model. Many different algorithms are available, you can find them under `quadra/experiment/generic/mnist/anomaly` folder.
+This will make use of the [anomalib](https://github.com/openvinotoolkit/anomalib) library to train the model. Many different algorithms are available, you can find them under `quadra/configs/experiment/generic/mnist/anomaly` folder.
 
 ## Running with Custom Datasets
 
-Each task comes with a default configuration file that can be customized for your needs. Each example experiment we have seen so far uses a default configuration file that can be found under `quadra/experiment/base/<task>/<config_name>.yaml`. 
+Each task comes with a default configuration file that can be customized for your needs. Each example experiment we have seen so far uses a default configuration file that can be found under `quadra/configs/experiment/base/<task>/<config_name>.yaml`. 
 
 Let's see how we can customize the configuration file to run the classification experiment on a custom dataset.
 

{quadra-2.2.1 → quadra-2.2.3}/README.md

@@ -172,7 +172,7 @@ On the same dataset we can run a simple SSL training using the BYOL algorithm wi
 quadra experiment=generic/imagenette/ssl/byol logger=csv
 ```
 
-BYOL is not the only SSL algorithm available, you can find a list of all the available algorithms under `quadra/experiment/generic/imagenette/ssl` folder.
+BYOL is not the only SSL algorithm available, you can find a list of all the available algorithms under `quadra/configs/experiment/generic/imagenette/ssl` folder.
 
 
 ### Anomaly Detection Training
@@ -185,11 +185,11 @@ quadra experiment=generic/mnist/anomaly/padim logger=csv
 
 This will run an anomaly detection considering on of the classes as good (default is the number 9) and the rest as anomalies.
 
-This will make use of the [anomalib](https://github.com/openvinotoolkit/anomalib) library to train the model. Many different algorithms are available, you can find them under `quadra/experiment/generic/mnist/anomaly` folder.
+This will make use of the [anomalib](https://github.com/openvinotoolkit/anomalib) library to train the model. Many different algorithms are available, you can find them under `quadra/configs/experiment/generic/mnist/anomaly` folder.
 
 ## Running with Custom Datasets
 
-Each task comes with a default configuration file that can be customized for your needs. Each example experiment we have seen so far uses a default configuration file that can be found under `quadra/experiment/base/<task>/<config_name>.yaml`. 
+Each task comes with a default configuration file that can be customized for your needs. Each example experiment we have seen so far uses a default configuration file that can be found under `quadra/configs/experiment/base/<task>/<config_name>.yaml`. 
 
 Let's see how we can customize the configuration file to run the classification experiment on a custom dataset.
 

{quadra-2.2.1 → quadra-2.2.3}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "quadra"
-version = "2.2.1"
+version = "2.2.3"
 description = "Deep Learning experiment orchestration library"
 authors = [
 	"Federico Belotti <federico.belotti@orobix.com>",
@@ -60,7 +60,7 @@ python_dotenv = "~0.21"
 rich = "~13.2"
 scikit_learn = "~1.2"
 pydantic = "1.10.10"
-grad_cam = "1.4.6"
+grad-cam-orobix = "1.5.3.dev001"
 matplotlib = "~3.6"
 seaborn = "~0.12"
 albumentations = "~1.3"
@@ -72,7 +72,7 @@ h5py = "~3.8"
 timm = "0.9.12"
 
 segmentation_models_pytorch-orobix = "0.3.3.dev1"
-anomalib-orobix = "0.7.0.dev140"
+anomalib-orobix = "0.7.0.dev141"
 xxhash = "~3.2"
 torchinfo = "~1.8"
 typing_extensions = { version = "4.11.0", python = "<3.10" }

{quadra-2.2.1 → quadra-2.2.3}/quadra/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "2.2.1"
+__version__ = "2.2.3"
 
 
 def get_version():

{quadra-2.2.1 → quadra-2.2.3}/quadra/main.py

@@ -1,6 +1,7 @@
 import time
 
 import hydra
+import matplotlib
 from omegaconf import DictConfig
 from pytorch_lightning import seed_everything
 
@@ -12,6 +13,8 @@ from quadra.utils.validator import validate_config
 load_envs()
 register_resolvers()
 
+
+matplotlib.use("Agg")
 log = get_logger(__name__)
 
 

{quadra-2.2.1 → quadra-2.2.3}/quadra/modules/classification/base.py

@@ -138,13 +138,9 @@ class ClassificationModule(BaseLightningModule):
         if isinstance(self.model.features_extractor, timm.models.resnet.ResNet):
             target_layers = [cast(BaseNetworkBuilder, self.model).features_extractor.layer4[-1]]
 
-            # Get model current device
-            device = next(self.model.parameters()).device
-
             self.cam = GradCAM(
                 model=self.model,
                 target_layers=target_layers,
-                use_cuda=device.type == "cuda",
             )
             # Activating gradients
             for p in self.model.features_extractor.layer4[-1].parameters():
@@ -262,7 +258,7 @@ class MultilabelClassificationModule(BaseLightningModule):
                 self.gradcam = False
             else:
                 target_layers = [model[0].features_extractor.layer4[-1]]
-                self.cam = GradCAM(model=model, target_layers=target_layers, use_cuda=torch.cuda.is_available())
+                self.cam = GradCAM(model=model, target_layers=target_layers)
 
     def forward(self, x):
         return self.model(x)

{quadra-2.2.1 → quadra-2.2.3}/quadra/tasks/classification.py

@@ -1138,7 +1138,6 @@ class ClassificationEvaluation(Evaluation[ClassificationDataModuleT]):
             self.cam = GradCAM(
                 model=self.deployment_model.model,
                 target_layers=target_layers,
-                use_cuda=(self.device != "cpu"),
             )
             for p in self.deployment_model.model.features_extractor.layer4[-1].parameters():
                 p.requires_grad = True

{quadra-2.2.1 → quadra-2.2.3}/quadra/utils/export.py

@@ -7,7 +7,6 @@ from typing import Any, Literal, TypeVar, cast
 import torch
 from anomalib.models.cflow import CflowLightning
 from omegaconf import DictConfig, ListConfig, OmegaConf
-from onnxconverter_common import auto_convert_mixed_precision
 from torch import nn
 
 from quadra.models.base import ModelSignatureWrapper
@@ -22,6 +21,7 @@ from quadra.utils.logger import get_logger
 try:
     import onnx  # noqa
     from onnxsim import simplify as onnx_simplify  # noqa
+    from onnxconverter_common import auto_convert_mixed_precision  # noqa
 
     ONNX_AVAILABLE = True
 except ImportError:
@@ -385,8 +385,9 @@ def _safe_export_half_precision_onnx(
             onnx.checker.check_model(onnx_model)
             return True
         except Exception as e:
-            log.debug("Failed to export model with mixed precision with error: %s", e)
-            return False
+            raise RuntimeError(
+                "Failed to export model with automatic mixed precision, check your model or disable ONNX export"
+            ) from e
     else:
         log.info("Exported half precision ONNX model does not contain NaN values, model is stable")
         return True

{quadra-2.2.1 → quadra-2.2.3}/quadra/utils/models.py

@@ -131,7 +131,6 @@ def get_feature(
             cam = GradCAM(
                 model=feature_extractor,
                 target_layers=target_layers,
-                use_cuda=torch.cuda.is_available(),
             )
             for p in feature_extractor.features_extractor.layer4[-1].parameters():
                 p.requires_grad = True