{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"nvidiaTeslaT4","dataSources":[{"sourceId":3649,"databundleVersionId":46718,"sourceType":"competition"},{"sourceId":302769,"sourceType":"modelInstanceVersion","isSourceIdPinned":true,"modelInstanceId":258507,"modelId":279727}],"dockerImageVersionId":30823,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"execution":{"iopub.status.busy":"2025-03-26T09:44:04.9578Z","iopub.execute_input":"2025-03-26T09:44:04.958214Z","iopub.status.idle":"2025-03-26T09:44:05.274763Z","shell.execute_reply.started":"2025-03-26T09:44:04.958185Z","shell.execute_reply":"2025-03-26T09:44:05.274004Z"},"jupyter":{"source_hidden":true}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T11:10:51.512249Z","iopub.execute_input":"2025-03-26T11:10:51.512585Z","iopub.status.idle":"2025-03-26T11:10:51.65376Z","shell.execute_reply.started":"2025-03-26T11:10:51.512561Z","shell.execute_reply":"2025-03-26T11:10:51.652911Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"!pip install -U \"huggingface_hub[cli]\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-27T13:16:28.526357Z","iopub.execute_input":"2025-03-27T13:16:28.526672Z","iopub.status.idle":"2025-03-27T13:16:35.018476Z","shell.execute_reply.started":"2025-03-27T13:16:28.526644Z","shell.execute_reply":"2025-03-27T13:16:35.017619Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"!pwd","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T10:04:20.337058Z","iopub.execute_input":"2025-03-26T10:04:20.337406Z","iopub.status.idle":"2025-03-26T10:04:20.456472Z","shell.execute_reply.started":"2025-03-26T10:04:20.337381Z","shell.execute_reply":"2025-03-26T10:04:20.45543Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"%ls","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T10:04:58.109386Z","iopub.execute_input":"2025-03-26T10:04:58.109707Z","iopub.status.idle":"2025-03-26T10:04:58.229463Z","shell.execute_reply.started":"2025-03-26T10:04:58.109683Z","shell.execute_reply":"2025-03-26T10:04:58.228656Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"!git clone https://github.com/Pham-Xuan-Thanh/PSLD.git\n! echo current dir $( pwd )\n%cd PSLD\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-27T13:16:35.019527Z","iopub.execute_input":"2025-03-27T13:16:35.01977Z","iopub.status.idle":"2025-03-27T13:16:35.929799Z","shell.execute_reply.started":"2025-03-27T13:16:35.019749Z","shell.execute_reply":"2025-03-27T13:16:35.928824Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"!huggingface-cli download thanhphxu/psld_cifar10_55m 'checkpoints/psld--epoch=2089-loss=0.0042.ckpt' --local-dir . --token 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\n# !huggingface-cli download thanhphxu/psld_cifar10_55m 'outputs/cifar10_test_images' --local-dir /kaggle/working/PSLD --token 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","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-27T13:16:37.679904Z","iopub.execute_input":"2025-03-27T13:16:37.680228Z","iopub.status.idle":"2025-03-27T13:16:50.285184Z","shell.execute_reply.started":"2025-03-27T13:16:37.680198Z","shell.execute_reply":"2025-03-27T13:16:50.284325Z"},"scrolled":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"!head checkpoints/psld--epoch=2089-loss=0.0042.ckpt","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-27T13:16:52.898975Z","iopub.execute_input":"2025-03-27T13:16:52.899272Z","iopub.status.idle":"2025-03-27T13:16:53.018747Z","shell.execute_reply.started":"2025-03-27T13:16:52.899248Z","shell.execute_reply":"2025-03-27T13:16:53.017708Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# !curl 'https://personalmicrosoftsoftware-my.sharepoint.com/personal/pandeyk1_personalmicrosoftsoftware_uci_edu/_layouts/15/download.aspx?UniqueId=91772ca7%2Db4e3%2D436c%2D9673%2D2f77650e1406' \\\n#   -H 'accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.7' \\\n#   -H 'accept-language: en-US,en;q=0.9' \\\n#   -H 'cache-control: no-cache' \\\n#   -b 'FeatureOverrides_experiments=[]; MSFPC=GUID=2601c284eb5d41c594a397d3c2dcd077&HASH=2601&LV=202409&V=4&LU=1726662138142; msal.cache.encryption=%7B%22id%22%3A%22019547e0-1024-7cf9-8ca7-140117d400cb%22%2C%22key%22%3A%22X0deg8AQE0mr_VJpyH-ISA6rAviI3A3G8SYHglZMm7k%22%7D; rtFa=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; SIMI=eyJzdCI6MH0=; MicrosoftApplicationsTelemetryDeviceId=4d0127c6-da94-4302-b9c3-cae851a4befd; ai_session=DHBrN1sEOfj4GsoOrfZNM/|1742982082883|1742982082883; nSGt-AD23666C9389064FC8B0D6B70C54B8A500A843977DA649C1=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; nSGt-480B957892B1D97D4518AE7153B9EE56C98885AFAB14742C=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; nSGt-BC2B439AC83E20E3E664C36BA943D8ACE8CBC311A1BDDFFC=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; nSGt-0905A3E275AD58857EEA61D2DACB1B8C3D0C1984466FDEB1=gYEwOEY4RkI1QTZBNzM1MzU1NUEzRDdERDY1NEM3QUE4NThFRDdFMTlCRDQ1MzBDREM2MDA5MDVBM0UyNzVBRDU4ODU3RUVBNjFEMkRBQ0IxQjhDM0QwQzE5ODQ0NjZGREVCMRIxMzM4NzQ1NTkzMDE4OTU5NzQrcGVyc29uYWxtaWNyb3NvZnRzb2Z0d2FyZS1teS5zaGFyZXBvaW50LmNvbWbfAO56ri4l53bjiDRWG+7V0PaHXSE1Dm7bUnxiP+HgZcyLS5V7UVXB8ge9LUY2egRvuoN4juRfY84hZ1izU38xo50o+0HoCGY+JbAa8iuSO7tVqSM8sdkdksp7IGpMehYlM0CfQ1fjD05ipBcmrYZBx+lMA4BDKwwg7czB5qEPYTwc1RCgGTZ0EQF8SR5O8F3z6NUa6eSdNbso/jNrUx2kPKksJRVNbosVrI7yeQybuhO7iMpqtAj8NuM1VACxdiiaHphlfKvcybyu+rQFOEd1w4I9tOqQcuSEzR5N4KqrrDpJJnIsi6Oq5aSbqdjWn3oj9DfUTeoThLZIjUuf4cGjAAAA; nSGt-AB8F05D24D09EBD75705FE254AFF585C020D917F9DE6B644=gYEwM0EwOUQxQTc3NUEyMkZFMjJBQ0Q4OUQ2RjFDRTJCNUY4REQ3MzIxNEQ4MjdCMEM1MEFCOEYwNUQyNEQwOUVCRDc1NzA1RkUyNTRBRkY1ODVDMDIwRDkxN0Y5REU2QjY0NBIxMzM4NzQ1NTkzMTQ4Mjk4NTkrcGVyc29uYWxtaWNyb3NvZnRzb2Z0d2FyZS1teS5zaGFyZXBvaW50LmNvbawQvr8W6c6OgSHexTf1wEtJEePUCWjYLgozUAJRjtlmOX51re543eWEL1u73WFIF1/N2Pw6UOEQv8fWXCG3lakM4GBylh5ZpQWgkjOpDxkSUPYOFfQbSFWPuMeH6WGYHV49PclXQA3reA+xcpPMLtoQpLK/IOJv7tgkSw1Ij2CJN9iGImQq8anRUod1Up03qIvfNFGRlew1Yuxg8F8oo9zr6AEoUABqSrey+7s7hbACHag12Et/u3bAMklhDFrjDWrBuZWDvfi6EzDyW3zWewk7hNuOYcH6Ws6X8HUEXF6ThWz+lYszRz4SwFNQfQIzi9RkP/GEWONvuOEs/KfSC8SjAAAA; nSGt-EA6FB8CC168942B4D8F94B25BA453EA9714C45F92DBB8478=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; RpsContextCookie=UHJldmlvdXNSZXF1ZXN0Q29ycmVsYXRpb25JZD0wZjQwOGVhMSUyRDcwMTklMkQ4MDAwJTJEN2Y4ZCUyRDE5NzcyYzc0OWI1NSZSZXR1cm5Vcmw9aHR0cHMlM0ElMkYlMkZwZXJzb25hbG1pY3Jvc29mdHNvZnR3YXJlJTJEbXklMkVzaGFyZXBvaW50JTJFY29tJTJGcGVyc29uYWwlMkZwYW5kZXlrMSU1RnBlcnNvbmFsbWljcm9zb2Z0c29mdHdhcmUlNUZ1Y2klNUZlZHUlMkYlNUZsYXlvdXRzJTJGMTUlMkZvbmVkcml2ZSUyRWFzcHglM0ZpZCUzRCUyRnBlcnNvbmFsJTJGcGFuZGV5azElNUZwZXJzb25hbG1pY3Jvc29mdHNvZnR3YXJlJTVGdWNpJTVGZWR1JTJGRG9jdW1lbnRzJTJGUFNMRCUyNmdhJTNEMSUyNm5vQXV0aFJlZGlyZWN0JTNEMQ==; FedAuth=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' \\\n#   -H 'pragma: no-cache' \\\n#   -H 'priority: u=0, i' \\\n#   -H 'referer: https://personalmicrosoftsoftware-my.sharepoint.com/personal/pandeyk1_personalmicrosoftsoftware_uci_edu/_layouts/15/onedrive.aspx?id=%2Fpersonal%2Fpandeyk1%5Fpersonalmicrosoftsoftware%5Fuci%5Fedu%2FDocuments%2FPSLD%2Fsota%2FCIFAR%2D10&ga=1' \\\n#   -H 'sec-ch-ua: \"Not(A:Brand\";v=\"99\", \"Google Chrome\";v=\"133\", \"Chromium\";v=\"133\"' \\\n#   -H 'sec-ch-ua-mobile: ?0' \\\n#   -H 'sec-ch-ua-platform: \"macOS\"' \\\n#   -H 'sec-fetch-dest: iframe' \\\n#   -H 'sec-fetch-mode: navigate' \\\n#   -H 'sec-fetch-site: same-origin' \\\n#   -H 'sec-fetch-user: ?1' \\\n#   -H 'service-worker-navigation-preload: {\"supportsFeatures\":[1855,61313]}' \\\n#   -H 'upgrade-insecure-requests: 1' \\\n#   -H 'user-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/133.0.0.0 Safari/537.36' \\\n#   -o cifar10_55m.ckpt ","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T09:45:26.442613Z","iopub.execute_input":"2025-03-26T09:45:26.443025Z","iopub.status.idle":"2025-03-26T09:45:27.44396Z","shell.execute_reply.started":"2025-03-26T09:45:26.442998Z","shell.execute_reply":"2025-03-26T09:45:27.443011Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# !conda create -n newCondaEnvironment python=3.8 -y\n# !source /opt/conda/bin/activate newCondaEnvironment && conda install python -y\n# !sudo rm /opt/conda/bin/python3\n# !sudo ln -sf /opt/conda/envs/newCondaEnvironment/bin/python3 /opt/conda/bin/python3","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T09:44:07.347128Z","iopub.execute_input":"2025-03-26T09:44:07.347465Z","iopub.status.idle":"2025-03-26T09:44:07.868618Z","shell.execute_reply.started":"2025-03-26T09:44:07.347409Z","shell.execute_reply":"2025-03-26T09:44:07.867458Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"\n\n# !pip install hydra-core --upgrade  pytorch-lightning==1.8.4   torchdiffeq   pipenv \"numpy<1.24\"   #torchvision  \n!pip install torch==1.13.1+cu116 -f https://download.pytorch.org/whl/torch_stable.html\n!pip install torchvision==0.14.1+cu116 -f https://download.pytorch.org/whl/torch_stable.html\n!pip install  torchdiffeq==0.2.3 torch-fidelity==0.3.0 torchmetrics==0.11.0  tqdm==4.64.1 wandb==0.13.5 scipy==1.9.3 scikit-learn==1.1.3 Pillow==9.3.0 numpy==1.23.5 matplotlib==3.6.2 hydra-core==1.2.0 pytorch-lightning==1.9.0 ninja==1.11.1\n# print(\"Installed packages. Continuing to load dependencies:\")\n# !pipenv install ","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-27T13:16:57.34669Z","iopub.execute_input":"2025-03-27T13:16:57.347004Z","iopub.status.idle":"2025-03-27T13:19:39.619204Z","shell.execute_reply.started":"2025-03-27T13:16:57.346979Z","shell.execute_reply":"2025-03-27T13:19:39.618291Z"},"scrolled":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Fine-tune","metadata":{}},{"cell_type":"code","source":"!python main/train_sde.py +dataset=cifar100/cifar100_psld \\\n     dataset.diffusion.data.root='/kaggle/working/PSLD/outputs' \\\n     dataset.diffusion.data.norm=True  \\\n     dataset.diffusion.data.hflip=True  \\\n     dataset.diffusion.model.score_fn.in_ch=6 \\\n     dataset.diffusion.model.score_fn.out_ch=6 \\\n     dataset.diffusion.model.score_fn.nf=128 \\\n     dataset.diffusion.model.score_fn.ch_mult=[2,2,2] \\\n     dataset.diffusion.model.score_fn.num_res_blocks=4 \\\n     dataset.diffusion.model.score_fn.attn_resolutions=[16] \\\n     dataset.diffusion.model.score_fn.dropout=0.15 \\\n     dataset.diffusion.model.score_fn.progressive_input='residual' \\\n     dataset.diffusion.model.score_fn.fir=True \\\n     dataset.diffusion.model.score_fn.embedding_type='fourier' \\\n     dataset.diffusion.model.sde.beta_min=8.0 \\\n     dataset.diffusion.model.sde.beta_max=8.0 \\\n     dataset.diffusion.model.sde.decomp_mode='lower' \\\n     dataset.diffusion.model.sde.nu=4.01 \\\n     dataset.diffusion.model.sde.gamma=0.01 \\\n     dataset.diffusion.model.sde.kappa=0.04 \\\n     dataset.diffusion.training.seed=0 \\\n     dataset.diffusion.training.mode='hsm' \\\n     dataset.diffusion.training.fp16=False \\\n     dataset.diffusion.training.use_ema=True \\\n     dataset.diffusion.training.batch_size=64 \\\n     dataset.diffusion.training.accelerator='gpu' \\\n     dataset.diffusion.training.workers=1 \\\n     dataset.diffusion.training.restore_path=\"/kaggle/working/PSLD/checkpoints/psld--epoch\\=2089-loss\\=0.0042.ckpt\" \\\n     dataset.diffusion.training.chkpt_interval=5 \\\n     dataset.diffusion.training.epochs=2101 \\\n     dataset.diffusion.training.devices=2 \\\n     dataset.diffusion.training.results_dir='./logs' \n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-27T13:19:39.620579Z","iopub.execute_input":"2025-03-27T13:19:39.620879Z","iopub.status.idle":"2025-03-27T13:21:32.778458Z","shell.execute_reply.started":"2025-03-27T13:19:39.620848Z","shell.execute_reply":"2025-03-27T13:21:32.777346Z"},"scrolled":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Inference ","metadata":{}},{"cell_type":"code","source":"# !python main/eval/sample.py +dataset=cifar10/cifar10_psld \\\n#                      dataset.diffusion.model.score_fn.in_ch=6 \\\n#                      dataset.diffusion.model.score_fn.out_ch=6 \\\n#                      dataset.diffusion.model.score_fn.nf=128 \\\n#                      dataset.diffusion.model.score_fn.ch_mult=[2,2,2] \\\n#                      dataset.diffusion.model.score_fn.num_res_blocks=8 \\\n#                      dataset.diffusion.model.score_fn.attn_resolutions=[16] \\\n#                      dataset.diffusion.model.score_fn.dropout=0.15 \\\n#                      dataset.diffusion.model.score_fn.progressive_input='residual' \\\n#                      dataset.diffusion.model.score_fn.fir=True \\\n#                      dataset.diffusion.model.score_fn.embedding_type='fourier' \\\n#                      dataset.diffusion.model.sde.beta_min=8.0 \\\n#                      dataset.diffusion.model.sde.beta_max=8.0 \\\n#                      dataset.diffusion.model.sde.nu=4.01 \\\n#                      dataset.diffusion.model.sde.gamma=0.01 \\\n#                      dataset.diffusion.model.sde.kappa=0.04 \\\n#                      dataset.diffusion.model.sde.decomp_mode='lower' \\\n#                      dataset.diffusion.evaluation.seed=0 \\\n#                      dataset.diffusion.evaluation.sample_prefix='gpu' \\\n#                      dataset.diffusion.evaluation.devices=2 \\\n#                      dataset.diffusion.evaluation.save_path='../outputs/' \\\n#                      dataset.diffusion.evaluation.batch_size=16 \\\n#                      dataset.diffusion.evaluation.stride_type='uniform' \\\n#                      dataset.diffusion.evaluation.sample_from='target' \\\n#                      dataset.diffusion.evaluation.workers=1 \\\n#                      dataset.diffusion.evaluation.chkpt_path='../cifar10_55m.ckpt' \\\n#                      dataset.diffusion.evaluation.sampler.name=\"em_sde\" \\\n#                      dataset.diffusion.evaluation.n_samples=50000 \\\n#                      dataset.diffusion.evaluation.n_discrete_steps=50 \\\n#                      dataset.diffusion.evaluation.path_prefix=\"50\"\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T09:44:47.325261Z","iopub.execute_input":"2025-03-26T09:44:47.325494Z","iopub.status.idle":"2025-03-26T09:44:47.423959Z","shell.execute_reply.started":"2025-03-26T09:44:47.325473Z","shell.execute_reply":"2025-03-26T09:44:47.422821Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Evaluate samples","metadata":{}},{"cell_type":"code","source":"# !python main/eval/sample.py +dataset=cifar100/cifar100_psld \\\n#                      dataset.diffusion.model.score_fn.in_ch=6 \\\n#                      dataset.diffusion.model.score_fn.out_ch=6 \\\n#                      dataset.diffusion.model.score_fn.nf=128 \\\n#                      dataset.diffusion.model.score_fn.ch_mult=[2,2,2] \\\n#                      dataset.diffusion.model.score_fn.num_res_blocks=4 \\\n#                      dataset.diffusion.model.score_fn.attn_resolutions=[16] \\\n#                      dataset.diffusion.model.score_fn.dropout=0.15 \\\n#                      dataset.diffusion.model.score_fn.progressive_input='Residual' \\\n#                      dataset.diffusion.model.score_fn.fir=True \\\n#                      dataset.diffusion.model.score_fn.embedding_type='Fourier' \\\n#                      dataset.diffusion.model.sde.beta_min=8.0 \\\n#                      dataset.diffusion.model.sde.beta_max=8.0 \\\n#                      dataset.diffusion.model.sde.nu=4.01 \\\n#                      dataset.diffusion.model.sde.gamma=0.01 \\\n#                      dataset.diffusion.model.sde.kappa=0.04 \\\n#                      dataset.diffusion.model.sde.decomp_mode='lower' \\\n#                      dataset.diffusion.evaluation.seed=0 \\\n#                      dataset.diffusion.evaluation.sample_prefix='test' \\\n#                      dataset.diffusion.evaluation.devices=2 \\\n#                      dataset.diffusion.evaluation.save_path=\"./outputs\" \\\n#                      dataset.diffusion.evaluation.batch_size=512 \\\n#                      dataset.diffusion.evaluation.stride_type='uniform' \\\n#                      dataset.diffusion.evaluation.sample_from='target' \\\n#                      dataset.diffusion.evaluation.workers=1 \\\n#                      dataset.diffusion.evaluation.chkpt_path=\"/kaggle/working/PSLD/psld_hsm_gamma\\=0.01_nu\\=4.01_cifar10_continuous_sfn\\=ncsnpp_nparams\\=55M.ckpt\" \\\n#                      dataset.diffusion.evaluation.sampler.name=\"em_sde\" \\\n#                      dataset.diffusion.evaluation.n_samples=50000 \\\n#                      dataset.diffusion.evaluation.n_discrete_steps=50 \\\n#                      dataset.diffusion.evaluation.path_prefix=\"50\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T10:19:44.311479Z","iopub.execute_input":"2025-03-26T10:19:44.31181Z","iopub.status.idle":"2025-03-26T10:20:17.039383Z","shell.execute_reply.started":"2025-03-26T10:19:44.311782Z","shell.execute_reply":"2025-03-26T10:20:17.03774Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Evaluate inpaint\n","metadata":{}},{"cell_type":"code","source":"# !python main/eval/sample.py +dataset=cifar10/cifar10_psld \\\n#                      dataset.diffusion.data.root='/kaggle/working/PSLD/outputs' \\\n#                      dataset.diffusion.data.name='cifar10' \\\n#                      dataset.diffusion.data.norm=True \\\n#                      dataset.diffusion.data.hflip=True \\\n#                      dataset.diffusion.data.return_target=True \\\n#                      dataset.diffusion.model.score_fn.in_ch=6 \\\n#                      dataset.diffusion.model.score_fn.out_ch=6 \\\n#                      dataset.diffusion.model.score_fn.nf=128 \\\n#                      dataset.diffusion.model.score_fn.ch_mult=[2,2,2] \\\n#                      dataset.diffusion.model.score_fn.num_res_blocks=4 \\\n#                      dataset.diffusion.model.score_fn.attn_resolutions=[16] \\\n#                      dataset.diffusion.model.score_fn.dropout=0.15 \\\n#                      dataset.diffusion.model.score_fn.progressive_input='Residual' \\\n#                      dataset.diffusion.model.score_fn.fir=True \\\n#                      dataset.diffusion.model.score_fn.embedding_type='Fourier' \\\n#                      dataset.diffusion.model.sde.beta_min=8.0 \\\n#                      dataset.diffusion.model.sde.beta_max=8.0 \\\n#                      dataset.diffusion.model.sde.nu=4.01 \\\n#                      dataset.diffusion.model.sde.gamma=0.01 \\\n#                      dataset.diffusion.model.sde.kappa=0.04 \\\n#                      dataset.diffusion.model.sde.decomp_mode='lower' \\\n#                      dataset.diffusion.evaluation.seed=0 \\\n#                      dataset.diffusion.evaluation.sample_prefix='test' \\\n#                      dataset.diffusion.evaluation.devices=2 \\\n#                      dataset.diffusion.evaluation.save_path=\"./outputs\" \\\n#                      dataset.diffusion.evaluation.batch_size=512 \\\n#                      dataset.diffusion.evaluation.stride_type='uniform' \\\n#                      dataset.diffusion.evaluation.sample_from='target' \\\n#                      dataset.diffusion.evaluation.workers=1 \\\n#                      dataset.diffusion.evaluation.chkpt_path=\"/kaggle/working/PSLD/psld_hsm_gamma\\=0.01_nu\\=4.01_cifar10_continuous_sfn\\=ncsnpp_nparams\\=55M.ckpt\" \\\n#                      dataset.diffusion.evaluation.sampler.name=\"em_sde\" \\\n#                      dataset.diffusion.evaluation.n_samples=5000 \\\n#                      dataset.diffusion.evaluation.n_discrete_steps=50 \\\n#                      dataset.diffusion.evaluation.path_prefix=\"50\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T10:52:53.991498Z","iopub.execute_input":"2025-03-26T10:52:53.991828Z","iopub.status.idle":"2025-03-26T11:10:12.346647Z","shell.execute_reply.started":"2025-03-26T10:52:53.991804Z","shell.execute_reply":"2025-03-26T11:10:12.3457Z"},"scrolled":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Benchmark","metadata":{}},{"cell_type":"code","source":"%cd /kaggle/working/PSLD\n!pwd","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T10:19:40.785116Z","iopub.execute_input":"2025-03-26T10:19:40.785459Z","iopub.status.idle":"2025-03-26T10:19:40.905712Z","shell.execute_reply.started":"2025-03-26T10:19:40.785424Z","shell.execute_reply":"2025-03-26T10:19:40.904922Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T11:28:30.354432Z","iopub.execute_input":"2025-03-26T11:28:30.354789Z","iopub.status.idle":"2025-03-26T11:30:06.53556Z","shell.execute_reply.started":"2025-03-26T11:28:30.354764Z","shell.execute_reply":"2025-03-26T11:30:06.534426Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"!pip install torch-fidelity","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T10:20:47.791002Z","iopub.execute_input":"2025-03-26T10:20:47.791333Z","iopub.status.idle":"2025-03-26T10:20:51.178821Z","shell.execute_reply.started":"2025-03-26T10:20:47.791309Z","shell.execute_reply":"2025-03-26T10:20:51.177922Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# !fidelity --gpu 0 --isc --input1 /kaggle/working/PSLD/outputs/2025-03-26/10-52-58/outputs/50/images","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T11:14:00.397076Z","iopub.execute_input":"2025-03-26T11:14:00.397422Z","iopub.status.idle":"2025-03-26T11:14:27.674635Z","shell.execute_reply.started":"2025-03-26T11:14:00.3974Z","shell.execute_reply":"2025-03-26T11:14:27.673805Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# !fidelity --gpu 0 --isc --input1 /kaggle/working/PSLD/outputs/2025-03-26/10-12-49/outputs/50/images","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T10:51:09.846864Z","iopub.execute_input":"2025-03-26T10:51:09.847288Z","iopub.status.idle":"2025-03-26T10:51:19.505509Z","shell.execute_reply.started":"2025-03-26T10:51:09.84725Z","shell.execute_reply":"2025-03-26T10:51:19.504668Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# !fidelity --gpu 0 --fid --input1 /kaggle/working/PSLD/outputs/cifar10_test_images --input2 /kaggle/working/PSLD/outputs/2025-03-26/10-52-58/outputs/50/images","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T11:25:54.428442Z","iopub.execute_input":"2025-03-26T11:25:54.428794Z","iopub.status.idle":"2025-03-26T11:27:11.609046Z","shell.execute_reply.started":"2025-03-26T11:25:54.428764Z","shell.execute_reply":"2025-03-26T11:27:11.607783Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# import os\n# import numpy as np\n# from torchvision import datasets\n# from PIL import Image\n# from tqdm import tqdm\n\n# # Set directory to save test images\n# save_dir = \"/kaggle/working/PSLD/outputs/cifar10_test_images\"\n# os.makedirs(save_dir, exist_ok=True)\n\n# # Download only the test dataset (train=False)\n# dataset = datasets.CIFAR10(root=\"./data\", train=False, download=True)\n# data, labels = dataset.data, dataset.targets  # Extract images and labels\n\n# # Save each test image as PNG\n# for idx, img_array in tqdm(enumerate(data), total=len(data), desc=\"Saving Test Images\"):\n#     img = Image.fromarray(img_array)  # Convert NumPy array to PIL image\n#     img.save(os.path.join(save_dir, f\"test_image_{idx}.png\"), format=\"PNG\")\n\n# print(f\"All test images saved in: {save_dir}\")\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T11:22:47.869332Z","iopub.execute_input":"2025-03-26T11:22:47.869722Z","iopub.status.idle":"2025-03-26T11:22:58.49827Z","shell.execute_reply.started":"2025-03-26T11:22:47.869696Z","shell.execute_reply":"2025-03-26T11:22:58.497512Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-03-26T11:16:21.908761Z","iopub.execute_input":"2025-03-26T11:16:21.909088Z","iopub.status.idle":"2025-03-26T11:16:22.030176Z","shell.execute_reply.started":"2025-03-26T11:16:21.909063Z","shell.execute_reply":"2025-03-26T11:16:22.029087Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ! fidelity --gpu 0 --fid \\","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}