make code public

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2023 Hamidreza Kamkari
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

configurations/ood/diffusions/degrade_denoise/cifar10_celeba_1.yaml

@@ -0,0 +1,46 @@
+base_model:
+  config_dir: checkpoints-znod8v3z/diffusion_cifar10_distinctions_aep6z5pr_final/run_config.json
+  checkpoint_dir: checkpoints-znod8v3z/diffusion_cifar10_distinctions_aep6z5pr_final/checkpoints/de_score-based-VP-diffusion_best_valid.pt
+data:
+  # specify the datasets and dataloader configurations for the in and out of distribution data.
+  in_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: cifar10
+      train_batch_size: 8
+      valid_batch_size: 8
+      test_batch_size: 8
+    pick_loader: train
+  out_of_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: celeba
+      train_batch_size: 8
+      valid_batch_size: 8
+      test_batch_size: 8
+    pick_loader: test
+ood:
+  # bypass the entire visualization process since there is no need to plot the histograms that take time!
+  bypass_visualization: True
+
+  # for reproducibility
+  seed: 100
+
+  use_dataloader: True
+  pick_count: 1
+
+  # The OOD detection method in use
+  method: ood.reconstruction.DegradeDenoise
+  method_args:
+    num_time_steps: 20
+    steps: 50
+    validation_size: 1
+    methods_to_include: ['l2', 'lpips']
+    verbose: 3
+    gamma: 1
+
+
+logger:
+  project: ood-detection-single-runs
+  entity: platypus-dgm
+  name: cifar10_vs_celeba_small_degrade_denoise_ood

configurations/ood/diffusions/degrade_denoise/cifar10_celeba_2.yaml

@@ -0,0 +1,46 @@
+base_model:
+  config_dir: checkpoints-znod8v3z/diffusion_cifar10_distinctions_aep6z5pr_final/run_config.json
+  checkpoint_dir: checkpoints-znod8v3z/diffusion_cifar10_distinctions_aep6z5pr_final/checkpoints/de_score-based-VP-diffusion_best_valid.pt
+data:
+  # specify the datasets and dataloader configurations for the in and out of distribution data.
+  in_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: cifar10
+      train_batch_size: 8
+      valid_batch_size: 8
+      test_batch_size: 8
+    pick_loader: train
+  out_of_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: cifar10
+      train_batch_size: 8
+      valid_batch_size: 8
+      test_batch_size: 8
+    pick_loader: test
+ood:
+  # bypass the entire visualization process since there is no need to plot the histograms that take time!
+  bypass_visualization: True
+
+  # for reproducibility
+  seed: 100
+
+  use_dataloader: True
+  pick_count: 1
+
+  # The OOD detection method in use
+  method: ood.reconstruction.DegradeDenoise
+  method_args:
+    num_time_steps: 20
+    steps: 50
+    validation_size: 1
+    methods_to_include: ['l2', 'lpips']
+    verbose: 3
+    gamma: 1
+
+
+logger:
+  project: ood-detection-single-runs
+  entity: platypus-dgm
+  name: cifar10_vs_celeba_small_degrade_denoise_in_distr

configurations/ood/diffusions/degrade_denoise/cifar10_svhn_1.yaml

@@ -0,0 +1,46 @@
+base_model:
+  config_dir: checkpoints-znod8v3z/diffusion_cifar10_distinctions_aep6z5pr_final/run_config.json
+  checkpoint_dir: checkpoints-znod8v3z/diffusion_cifar10_distinctions_aep6z5pr_final/checkpoints/de_score-based-VP-diffusion_best_valid.pt
+data:
+  # specify the datasets and dataloader configurations for the in and out of distribution data.
+  in_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: cifar10
+      train_batch_size: 8
+      valid_batch_size: 8
+      test_batch_size: 8
+    pick_loader: train
+  out_of_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: svhn
+      train_batch_size: 8
+      valid_batch_size: 8
+      test_batch_size: 8
+    pick_loader: test
+ood:
+  # bypass the entire visualization process since there is no need to plot the histograms that take time!
+  bypass_visualization: True
+
+  # for reproducibility
+  seed: 100
+
+  use_dataloader: True
+  pick_count: 1
+
+  # The OOD detection method in use
+  method: ood.reconstruction.DegradeDenoise
+  method_args:
+    num_time_steps: 10
+    steps: 100
+    validation_size: 1
+    methods_to_include: ['l2', 'lpips']
+    verbose: 3
+    gamma: 1
+
+
+logger:
+  project: ood-detection-single-runs
+  entity: platypus-dgm
+  name: cifar10_vs_svhn_ood_degrade_denoise_ood

configurations/ood/diffusions/degrade_denoise/cifar10_svhn_2.yaml

@@ -0,0 +1,46 @@
+base_model:
+  config_dir: checkpoints-znod8v3z/diffusion_cifar10_distinctions_aep6z5pr_final/run_config.json
+  checkpoint_dir: checkpoints-znod8v3z/diffusion_cifar10_distinctions_aep6z5pr_final/checkpoints/de_score-based-VP-diffusion_best_valid.pt
+data:
+  # specify the datasets and dataloader configurations for the in and out of distribution data.
+  in_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: cifar10
+      train_batch_size: 8
+      valid_batch_size: 8
+      test_batch_size: 8
+    pick_loader: train
+  out_of_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: cifar10
+      train_batch_size: 8
+      valid_batch_size: 8
+      test_batch_size: 8
+    pick_loader: test
+ood:
+  # bypass the entire visualization process since there is no need to plot the histograms that take time!
+  bypass_visualization: True
+
+  # for reproducibility
+  seed: 100
+
+  use_dataloader: True
+  pick_count: 1
+
+  # The OOD detection method in use
+  method: ood.reconstruction.DegradeDenoise
+  method_args:
+    num_time_steps: 10
+    steps: 100
+    validation_size: 1
+    methods_to_include: ['l2', 'lpips']
+    verbose: 3
+    gamma: 1
+
+
+logger:
+  project: ood-detection-single-runs
+  entity: platypus-dgm
+  name: cifar10_vs_svhn_ood_degrade_denoise_in_distr

configurations/ood/diffusions/degrade_denoise/fmnist_mnist_1.yaml

@@ -0,0 +1,49 @@
+base_model:
+  config_dir: checkpoints-hg9x5r0n/diffusion_fashion-mnist_interparental_bxnuw7zk_final/run_config.json
+  checkpoint_dir: checkpoints-hg9x5r0n/diffusion_fashion-mnist_interparental_bxnuw7zk_final/checkpoints/de_score-based-VP-diffusion_best_valid.pt
+data:
+  # specify the datasets and dataloader configurations for the in and out of distribution data.
+  in_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: fashion-mnist
+      train_batch_size: 128
+      valid_batch_size: 128
+      test_batch_size: 128
+      additional_dataset_args:
+        resize_image: [32, 32]
+    pick_loader: train
+  out_of_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: mnist
+      train_batch_size: 128
+      valid_batch_size: 128
+      test_batch_size: 128
+      additional_dataset_args:
+        resize_image: [32, 32]
+    pick_loader: test
+ood:
+  # bypass the entire visualization process since there is no need to plot the histograms that take time!
+  bypass_visualization: True
+
+  # for reproducibility
+  seed: 100
+
+  use_dataloader: True
+  pick_count: 1
+
+  # The OOD detection method in use
+  method: ood.reconstruction.DegradeDenoise
+  method_args:
+    num_time_steps: 10
+    steps: 100
+    validation_size: 1
+    methods_to_include: ['l2', 'lpips']
+    verbose: 3
+
+
+logger:
+  project: ood-detection-single-runs
+  entity: platypus-dgm
+  name: fmnist_vs_mnist_degrade_denoise_ood

configurations/ood/diffusions/degrade_denoise/fmnist_mnist_2.yaml

@@ -0,0 +1,49 @@
+base_model:
+  config_dir: checkpoints-hg9x5r0n/diffusion_fashion-mnist_interparental_bxnuw7zk_final/run_config.json
+  checkpoint_dir: checkpoints-hg9x5r0n/diffusion_fashion-mnist_interparental_bxnuw7zk_final/checkpoints/de_score-based-VP-diffusion_best_valid.pt
+data:
+  # specify the datasets and dataloader configurations for the in and out of distribution data.
+  in_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: fashion-mnist
+      train_batch_size: 128
+      valid_batch_size: 128
+      test_batch_size: 128
+      additional_dataset_args:
+        resize_image: [32, 32]
+    pick_loader: train
+  out_of_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: fashion-mnist
+      train_batch_size: 128
+      valid_batch_size: 128
+      test_batch_size: 128
+      additional_dataset_args:
+        resize_image: [32, 32]
+    pick_loader: test
+ood:
+  # bypass the entire visualization process since there is no need to plot the histograms that take time!
+  bypass_visualization: True
+
+  # for reproducibility
+  seed: 100
+
+  use_dataloader: True
+  pick_count: 1
+
+  # The OOD detection method in use
+  method: ood.reconstruction.DegradeDenoise
+  method_args:
+    num_time_steps: 10
+    steps: 100
+    validation_size: 1
+    methods_to_include: ['l2', 'lpips']
+    verbose: 3
+
+
+logger:
+  project: ood-detection-single-runs
+  entity: platypus-dgm
+  name: fmnist_vs_mnist_degrade_denoise_in_distr

configurations/ood/diffusions/degrade_denoise/fmnist_mnist_generated_1.yaml

@@ -0,0 +1,54 @@
+base_model:
+  config_dir: checkpoints-hg9x5r0n/diffusion_fashion-mnist_interparental_bxnuw7zk_final/run_config.json
+  checkpoint_dir: checkpoints-hg9x5r0n/diffusion_fashion-mnist_interparental_bxnuw7zk_final/checkpoints/de_score-based-VP-diffusion_best_valid.pt
+data:
+  # specify the datasets and dataloader configurations for the in and out of distribution data.
+  in_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      dataset: fashion-mnist
+      train_batch_size: 128
+      valid_batch_size: 128
+      test_batch_size: 128
+      additional_dataset_args:
+        resize_image: [32, 32]
+    pick_loader: train
+  out_of_distribution:
+    dataloader_args:
+      make_valid_loader: false
+      train_batch_size: 128
+      valid_batch_size: 128
+      test_batch_size: 128
+      dataset: dgm-generated
+      dgm_args:
+        model_loading_config:
+          config_dir: checkpoints-hg9x5r0n/diffusion_fashion-mnist_interparental_bxnuw7zk_final/run_config.json
+          checkpoint_dir: checkpoints-hg9x5r0n/diffusion_fashion-mnist_interparental_bxnuw7zk_final/checkpoints/de_score-based-VP-diffusion_best_valid.pt
+        seed: 10
+        length: 1000
+    pick_loader: test
+ood:
+  # bypass the entire visualization process since there is no need to plot the histograms that take time!
+  bypass_visualization: True
+  histogram_limit: 0
+
+  # for reproducibility
+  seed: 100
+
+  use_dataloader: True
+  pick_count: 1
+
+  # The OOD detection method in use
+  method: ood.reconstruction.DegradeDenoise
+  method_args:
+    num_time_steps: 10
+    steps: 100
+    validation_size: 1
+    methods_to_include: ['l2', 'lpips']
+    verbose: 3
+
+
+logger:
+  project: ood-detection-single-runs
+  entity: platypus-dgm
+  name: fmnist_vs_mnist_degrade_denoise_ood