{"metadata":{"kaggle":{"accelerator":"nvidiaTeslaT4","dataSources":[{"sourceId":22990,"databundleVersionId":2048213,"sourceType":"competition"},{"sourceId":7032575,"sourceType":"datasetVersion","datasetId":4045246},{"sourceId":7037091,"sourceType":"datasetVersion","datasetId":4045253}],"dockerImageVersionId":30558,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true},"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"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"},"papermill":{"default_parameters":{},"duration":1779.497699,"end_time":"2023-11-24T00:37:46.857056","environment_variables":{},"exception":null,"input_path":"__notebook__.ipynb","output_path":"__notebook__.ipynb","parameters":{},"start_time":"2023-11-24T00:08:07.359357","version":"2.4.0"},"widgets":{"application/vnd.jupyter.widget-state+json":{"state":{"00045c7684d7400f81c4d539110892d6":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_ce31758adbf54f5b816a29498bb2989a","placeholder":"​","style":"IPY_MODEL_e017bd0ce2274f56acc5d5eff1b4b319","value":" 47/47 [06:34<00:00, 5.78s/it]"}},"02993fa460ae472abbbfb2f8c14422ec":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"0a72ccca7cca4609b376c260a46d9a3d":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_b2865f7678104b64a488a09adaadbd03","placeholder":"​","style":"IPY_MODEL_715172aad8c24e7c81fa1af332f93568","value":" 58/58 [02:06<00:00, 1.18it/s]"}},"0dd872c7a8b6472da955bd5c8206b457":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"0eb8084407ae4a0aa554fad7ba97ab57":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_132015c5572b468ca0b1ce2a95ed0779","IPY_MODEL_b574a2c40e054621852e46a800cec3c1","IPY_MODEL_d88dc7424c444696907d6ce92766c0b7"],"layout":"IPY_MODEL_02993fa460ae472abbbfb2f8c14422ec"}},"0ef323ab50e3417693cd7d7a63e99446":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"132015c5572b468ca0b1ce2a95ed0779":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_5c025219efd340a990395ab4579871f6","placeholder":"​","style":"IPY_MODEL_212fed197e744badab847408b464309a","value":"Segmenting Image: 100%"}},"1750eb66a3b94e979635e2b4b79bd260":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"1b561aadb6f345a8b6b78bf816951c1b":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_5000713e85ca4a33a2d1324a3b346bc2","max":1,"min":0,"orientation":"horizontal","style":"IPY_MODEL_2edfebd91e15439e8079c41022680ec7","value":1}},"1c3eddf50c92406d92b5e90ecd5d92e7":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_859a0527754944819a3eba842bc5432f","placeholder":"​","style":"IPY_MODEL_476dfb3d5a8e4660957bdc6fffd738c9","value":"Segmenting Image: 100%"}},"1f4c9f9dfbb140f592c780d3a3c77592":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_e6b0be6ec376407db021b33981c25b0c","placeholder":"​","style":"IPY_MODEL_c5dbeccd6e6b41e78e54fa53499769d0","value":" 92/92 [03:44<00:00, 1.17it/s]"}},"20395e64d51d44ee9c9e032e08ee1e4f":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_a15a75bd6ea747b6a8ba7b415615914d","placeholder":"​","style":"IPY_MODEL_0dd872c7a8b6472da955bd5c8206b457","value":"Segmenting Image: 100%"}},"212fed197e744badab847408b464309a":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"21d5c917d082419599b221ef4741cc37":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"228f6d24f1f349289aecf08043d5b9c9":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_b80f698eae3e47a2a88a159edaef11f2","max":47,"min":0,"orientation":"horizontal","style":"IPY_MODEL_3061322ba19345a78d1913e35588d66c","value":47}},"256f6b67d7684535bb28a62944ca7783":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"28acedf33d704fc89db4a2e7469510a8":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_50bb82b1e70e45e3875fea433142df3c","IPY_MODEL_1b561aadb6f345a8b6b78bf816951c1b","IPY_MODEL_393b607f85c84f9fa9b13c94d9609455"],"layout":"IPY_MODEL_5eb8f60deb5c4f0e8dd2ef22eed4d717"}},"2b9bc10d476047df9797676bb043716c":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"2bdcf9742856470ab8bd9c7f25d731b8":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"2d0a75352fba466a8187dd17b1258b3e":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_b04e20b0455148879ecf0ff4176622e5","IPY_MODEL_7a0e3f37064949ffab29f2eae4c00ca5","IPY_MODEL_449b580f60c54024b3fbed30de8e7a32"],"layout":"IPY_MODEL_3678f23794334f55a15d6ca95e6d4d36"}},"2edfebd91e15439e8079c41022680ec7":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"2fc43aef62ac4deda561ba0f0c63625e":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_20395e64d51d44ee9c9e032e08ee1e4f","IPY_MODEL_9abbe310c7ff4983abaa8b7eaecd87f2","IPY_MODEL_1f4c9f9dfbb140f592c780d3a3c77592"],"layout":"IPY_MODEL_d96b427cd7a64aee9b44b3a1f2fd1a3e"}},"3061322ba19345a78d1913e35588d66c":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"3235dde6c7a74a199c4398245c7d476e":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"3678f23794334f55a15d6ca95e6d4d36":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"393b607f85c84f9fa9b13c94d9609455":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_2b9bc10d476047df9797676bb043716c","placeholder":"​","style":"IPY_MODEL_3235dde6c7a74a199c4398245c7d476e","value":" 5/? [29:25<00:00, 365.71s/it]"}},"3d5ab9e9464548c2a59d1f212ac063d8":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"425624160e82477db9bdadcda1cf41a6":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"449b580f60c54024b3fbed30de8e7a32":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_b2a5f88f89a54484b489770093ed6961","placeholder":"​","style":"IPY_MODEL_2bdcf9742856470ab8bd9c7f25d731b8","value":" 60/60 [08:07<00:00, 4.35s/it]"}},"476dfb3d5a8e4660957bdc6fffd738c9":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"4c6d9510e7df49e3b0025bb5711ccb0f":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"5000713e85ca4a33a2d1324a3b346bc2":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":"20px"}},"50bb82b1e70e45e3875fea433142df3c":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_256f6b67d7684535bb28a62944ca7783","placeholder":"​","style":"IPY_MODEL_d4c403e951a34490bd6033c9fca2e5a7","value":""}},"5308be5fe2d247d6bff364075ee5f0a9":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"55e94d0cf0ab420cbc5e30fcb797e79b":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"5ad8425eab8249f0a2b8c383feed7d47":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"5c025219efd340a990395ab4579871f6":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"5eb8f60deb5c4f0e8dd2ef22eed4d717":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"6c920a3c2aaf49f190f34d71f922bf4c":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"715172aad8c24e7c81fa1af332f93568":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"7a0e3f37064949ffab29f2eae4c00ca5":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_91c105de160443d9951f30c2bdadd7c3","max":60,"min":0,"orientation":"horizontal","style":"IPY_MODEL_ac8e179cf93f4ef2afb4d43f84b3c05e","value":60}},"826b6c5c96f7440fa4951da428ecb729":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"859a0527754944819a3eba842bc5432f":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"91c105de160443d9951f30c2bdadd7c3":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"9abbe310c7ff4983abaa8b7eaecd87f2":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_21d5c917d082419599b221ef4741cc37","max":92,"min":0,"orientation":"horizontal","style":"IPY_MODEL_1750eb66a3b94e979635e2b4b79bd260","value":92}},"9dd7e0dc36654aeca1ea6a70e95e1d7e":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_e7437d191cbb4d8ab7cb7d411929e90c","IPY_MODEL_c611b115bf0340369dea8183132dce1d","IPY_MODEL_0a72ccca7cca4609b376c260a46d9a3d"],"layout":"IPY_MODEL_e9d410c9ac294a1783f49de9c1a539f2"}},"a15a75bd6ea747b6a8ba7b415615914d":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"ac8e179cf93f4ef2afb4d43f84b3c05e":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"b04e20b0455148879ecf0ff4176622e5":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_e780eeb1da604a53b04f4e4d84ea088f","placeholder":"​","style":"IPY_MODEL_5ad8425eab8249f0a2b8c383feed7d47","value":"Segmenting Image: 100%"}},"b2865f7678104b64a488a09adaadbd03":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"b2a5f88f89a54484b489770093ed6961":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"b574a2c40e054621852e46a800cec3c1":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_5308be5fe2d247d6bff364075ee5f0a9","max":65,"min":0,"orientation":"horizontal","style":"IPY_MODEL_0ef323ab50e3417693cd7d7a63e99446","value":65}},"b80f698eae3e47a2a88a159edaef11f2":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"c475d11be383425c8d846c9a405d5c5e":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_1c3eddf50c92406d92b5e90ecd5d92e7","IPY_MODEL_228f6d24f1f349289aecf08043d5b9c9","IPY_MODEL_00045c7684d7400f81c4d539110892d6"],"layout":"IPY_MODEL_826b6c5c96f7440fa4951da428ecb729"}},"c5dbeccd6e6b41e78e54fa53499769d0":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"c611b115bf0340369dea8183132dce1d":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_6c920a3c2aaf49f190f34d71f922bf4c","max":58,"min":0,"orientation":"horizontal","style":"IPY_MODEL_3d5ab9e9464548c2a59d1f212ac063d8","value":58}},"ce31758adbf54f5b816a29498bb2989a":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"d4c403e951a34490bd6033c9fca2e5a7":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"d88dc7424c444696907d6ce92766c0b7":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_425624160e82477db9bdadcda1cf41a6","placeholder":"​","style":"IPY_MODEL_e20656a653bc4555a1b669f3ddc43aeb","value":" 65/65 [04:01<00:00, 3.63s/it]"}},"d96b427cd7a64aee9b44b3a1f2fd1a3e":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"e017bd0ce2274f56acc5d5eff1b4b319":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"e20656a653bc4555a1b669f3ddc43aeb":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"DescriptionStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"e6b0be6ec376407db021b33981c25b0c":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"e7437d191cbb4d8ab7cb7d411929e90c":{"model_module":"@jupyter-widgets/controls","model_module_version":"1.5.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_4c6d9510e7df49e3b0025bb5711ccb0f","placeholder":"​","style":"IPY_MODEL_55e94d0cf0ab420cbc5e30fcb797e79b","value":"Segmenting Image: 100%"}},"e780eeb1da604a53b04f4e4d84ea088f":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"e9d410c9ac294a1783f49de9c1a539f2":{"model_module":"@jupyter-widgets/base","model_module_version":"1.2.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}}},"version_major":2,"version_minor":0}}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"# Leave lines below uncommented this if you get:\n# OMP: Error #15: Initializing libiomp5md.dll, but found libiomp5md.dll already initialized.\nimport os\nos.environ['KMP_DUPLICATE_LIB_OK']='True'\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)","metadata":{"_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","papermill":{"duration":0.866671,"end_time":"2023-11-24T00:08:11.847917","exception":false,"start_time":"2023-11-24T00:08:10.981246","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:32:21.533861Z","iopub.execute_input":"2023-11-24T14:32:21.534270Z","iopub.status.idle":"2023-11-24T14:32:21.897545Z","shell.execute_reply.started":"2023-11-24T14:32:21.534237Z","shell.execute_reply":"2023-11-24T14:32:21.896548Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import torch\nimport torch.nn as nn\nfrom torch.utils.data import DataLoader\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom matplotlib.patches import Polygon\nfrom tqdm import tqdm\nimport gc\nimport json\nimport cv2\nimport tifffile as tiff\nfrom skimage import color","metadata":{"papermill":{"duration":4.510089,"end_time":"2023-11-24T00:08:16.363021","exception":false,"start_time":"2023-11-24T00:08:11.852932","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:32:21.899450Z","iopub.execute_input":"2023-11-24T14:32:21.899968Z","iopub.status.idle":"2023-11-24T14:32:23.540054Z","shell.execute_reply.started":"2023-11-24T14:32:21.899931Z","shell.execute_reply":"2023-11-24T14:32:23.539050Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import sys\n\n# Import functions\nsys.path.insert(0, os.path.join(os.getcwd(), '..', 'input', 'modules'))\nfrom glomerulus import Glomerulus, Patch, KidneySampleDataset, get_glomeruli, generate_glomerulus_patches\nfrom networks import UNet_Large, CBAM_R2UNet_v2_Large\n# from utils import read_tiff","metadata":{"papermill":{"duration":0.50199,"end_time":"2023-11-24T00:08:16.870103","exception":false,"start_time":"2023-11-24T00:08:16.368113","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:32:23.541489Z","iopub.execute_input":"2023-11-24T14:32:23.541955Z","iopub.status.idle":"2023-11-24T14:32:23.837770Z","shell.execute_reply.started":"2023-11-24T14:32:23.541903Z","shell.execute_reply":"2023-11-24T14:32:23.836906Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def read_tiff(image_path):\n with rasterio.open(image_path) as image:\n image_data = image.read().astype(np.float32)\n del image\n gc.collect()\n print(image_data.dtype, image_data.shape)\n \n if image_data.shape[0] == 1:\n if image_data.any() > 1:\n image_data /= 255.0\n \n image_data = image_data[0]\n gc.collect()\n else:\n if image_data.any() > 1:\n image_data /= 255.0\n \n if image_data.shape[0] == 3:\n image_data = np.transpose(image_data, (1, 2, 0))\n image_data = color.rgb2gray(image_data)\n \n print(image_data.shape, image_data.dtype)\n gc.collect()\n \n return image_data","metadata":{"papermill":{"duration":0.014011,"end_time":"2023-11-24T00:08:16.889041","exception":false,"start_time":"2023-11-24T00:08:16.875030","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:32:23.839926Z","iopub.execute_input":"2023-11-24T14:32:23.840398Z","iopub.status.idle":"2023-11-24T14:32:23.848750Z","shell.execute_reply.started":"2023-11-24T14:32:23.840368Z","shell.execute_reply":"2023-11-24T14:32:23.847396Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Define paths\nbase_dir = os.path.join(os.getcwd(), '..', 'input')\n\n# Images\nimage_names = [\n 'afa5e8098',\n '4ef6695ce',\n 'c68fe75ea',\n '26dc41664',\n '095bf7a1f',\n '54f2eec69',\n '1e2425f28',\n 'e79de561c',\n 'cb2d976f4',\n 'b9a3865fc',\n '8242609fa',\n '0486052bb',\n '2f6ecfcdf',\n 'b2dc8411c',\n 'aaa6a05cc'\n]\ntrain_images_dir = os.path.join(base_dir, 'hubmap-kidney-segmentation/train')\n\ntest_image_path = os.path.join(train_images_dir, '26dc41664.tiff')\ntest_label_path = os.path.join(train_images_dir, '26dc41664.json')\n\n# Models\nmodels_dir = os.path.join(base_dir, 'pretrained')\nunet_path = os.path.join(models_dir, 'UNet_Large_BCE_Dice_0.6_0.4_5000_16_100_best_loss.npz')\ncbam_r2unet_path = os.path.join(models_dir, 'CBAM_R2UNet_v2_Large_BCE_Dice_0.6_0.4_5000_16_100_best_loss.npz')","metadata":{"papermill":{"duration":0.013941,"end_time":"2023-11-24T00:08:16.907455","exception":false,"start_time":"2023-11-24T00:08:16.893514","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:32:23.850056Z","iopub.execute_input":"2023-11-24T14:32:23.850365Z","iopub.status.idle":"2023-11-24T14:32:23.867637Z","shell.execute_reply.started":"2023-11-24T14:32:23.850340Z","shell.execute_reply":"2023-11-24T14:32:23.866840Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def test(model,test_image,test_glomeruli,image_value_lb = 100,image_value_ub = 205):\n out_img = np.zeros((\n int(np.ceil(test_image.shape[1] / 512) * 512),\n int(np.ceil(test_image.shape[2] / 512) * 512)\n ))\n intersection = 0\n union = 0\n smooth = 0.00000001\n for idx_y in tqdm(range(0, test_image.shape[1], 512)):\n for idx_x in range(0, test_image.shape[2], 512):\n center_x = idx_x + 512 // 2\n center_y = idx_y + 512 // 2\n test_patch = Patch(\n center_x = center_x,\n center_y = center_y,\n theta = 0,\n patch_size = 512,\n glomeruli = test_glomeruli,\n image = test_image\n )\n image = test_patch.render_image()\n label = test_patch.render_mask()\n mean = np.mean(image[0])\n if mean >image_value_ub or mean image_value_ub or mean image_value_ub or mean image_value_ub or mean 0:\n print(patches)\n # Create dataset and dataloader from patches\n patches_dataset = KidneySampleDataset(patches)\n\n if windows:\n patches_dataloader = DataLoader(patches_dataset, batch_size=batch_size, shuffle=False)\n else:\n patches_dataloader = DataLoader(patches_dataset, batch_size=batch_size, shuffle=False, num_workers=num_workers, prefetch_factor=prefetch_factor)\n\n # Run model\n model.eval()\n idx = 0\n with torch.no_grad():\n for inputs, labels in patches_dataloader:\n # Clear gpu cache\n gc.collect()\n torch.cuda.empty_cache()\n # Move data to GPU\n inputs = torch.unsqueeze(torch.tensor(inputs.type(torch.cuda.FloatTensor).cuda()), axis=1)\n # Run model\n outputs = model.forward(inputs)\n\n out_array = torch.squeeze(outputs, axis=1).cpu().detach().numpy()\n\n # Loops through model output and transfer onto output image array\n for patch in out_array:\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n out_img[idx_y: idx_y+512, idx_x: idx_x+512] += patch\n idx += 1\n return out_img\ndef test_stride_dataloader_v3( \n model, \n test_image, \n image_value_lb = 100, \n image_value_ub = 205, \n stride = 100, \n windows = False, \n batch_size = 16, \n num_workers = 8, \n prefetch_factor = 1 \n ): \n \n out_img = np.zeros(( \n int(np.ceil(test_image.shape[1] / 512) * 512), \n int(np.ceil(test_image.shape[2] / 512) * 512) \n )) \n out_img = np.where(out_img >= 0.5, 1, 0).astype(np.uint8)\n patches = [] \n patch_coordinates = [] \n for idx_y in tqdm(range(0, test_image.shape[1], stride), desc='Segmenting Image'): \n for idx_x in range(0, test_image.shape[2], stride): \n center_x = idx_x + 512 // 2 \n center_y = idx_y + 512 // 2 \n test_patch = Patch( \n center_x = center_x, \n center_y = center_y, \n theta = 0, \n patch_size = 512, \n glomeruli = None, \n image = test_image \n ) \n \n image = test_patch.render_image() \n mean = np.mean(image[0]) \n del image\n if mean >image_value_ub or mean 0.5,1,0).astype(np.uint8)\n out_img[idx_y: idx_y+512, idx_x: idx_x+512] += patch \n idx += 1 \n del patch\n del inputs\n del patches_dataloader\n del patch_coordinates\n del patches\n del patches_dataset\n gc.collect()\n return out_img\n\ndef test_stride_dataloader_v4(\n model1,\n model2,\n test_image,\n image_value_lb = 100,\n image_value_ub = 205, \n stride = 100,\n windows = False,\n batch_size = 16,\n num_workers = 8,\n prefetch_factor = 1,\n size_lowerbound = 10000,\n ):\n \n # Initialize array for output image\n out_img = np.zeros((test_image.shape[1],test_image.shape[2]), dtype=np.uint8)\n \n \n patches = []\n patch_coordinates = []\n \n # Segment input image into patches and storing into list\n for idx_y in tqdm(range(0, test_image.shape[1], stride), desc='Segmenting Image'):\n for idx_x in range(0, test_image.shape[2], stride):\n \n # Generate patch object\n center_x = idx_x + 512 // 2\n center_y = idx_y + 512 // 2\n test_patch = Patch(\n center_x = center_x,\n center_y = center_y,\n theta = 0,\n patch_size = 512,\n glomeruli = None,\n image = test_image\n )\n \n # Render patch image and test if kidney sample is present through thresholding\n # Too dark / light means the patch is empty\n image = test_patch.render_image()\n mean = np.mean(image[0])\n if mean >image_value_ub or mean = 16:\n # Create dataset and dataloader from patches\n patches_dataset = KidneySampleDataset(patches)\n if windows:\n patches_dataloader = DataLoader(patches_dataset, batch_size=batch_size, shuffle=False)\n else:\n patches_dataloader = DataLoader(patches_dataset, batch_size=batch_size, shuffle=False, num_workers=num_workers, prefetch_factor=prefetch_factor)\n\n # Run model\n model1.eval()\n model2.eval()\n idx = 0\n with torch.no_grad():\n for inputs, labels in patches_dataloader:\n # Clear gpu cache\n gc.collect()\n torch.cuda.empty_cache()\n # Move data to GPU\n inputs = torch.unsqueeze(torch.tensor(inputs.type(torch.cuda.FloatTensor)), axis=1)\n \n # Run and get outputs from both models\n outputs1 = model1.forward(inputs.cuda(device=0))\n outputs2 = model2.forward(inputs.cuda(device=0))\n del inputs\n \n # Transfer output tensors from gpus and convert to numpy\n out_array1 = torch.squeeze(outputs1, axis=1).cpu().detach().numpy()\n out_array2 = torch.squeeze(outputs2, axis=1).cpu().detach().numpy()\n del outputs1\n del outputs2\n \n # === Post processing model1 output ===\n \n prev_idx = idx\n # Loops through model output and transfer onto output image array\n for patch in out_array1:\n # Convert to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n \n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n\n max_y = min(idx_y+512, test_image.shape[1])\n max_x = min(idx_x+512, test_image.shape[2])\n \n out_img[idx_y: max_y, idx_x: max_x] += patch[:max_y - idx_y, :max_x - idx_x]\n idx += 1\n del out_array1\n \n # === Post processing model2 output ===\n \n idx = prev_idx\n # Loops through model output and transfer onto output image array\n for patch in out_array2:\n # Converty to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n \n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n \n max_y = min(idx_y+512, test_image.shape[1])\n max_x = min(idx_x+512, test_image.shape[2])\n \n out_img[idx_y: max_y, idx_x: max_x] = np.where(out_img[idx_y: max_y, idx_x: max_x] + patch[:max_y - idx_y, :max_x - idx_x] >= 1, 1, 0)\n idx += 1\n del out_array2\n del patches\n del patch_coordinates\n del patches_dataloader\n \n patches = []\n patch_coordinates = []\n \n # Process last patches if number of patches isn't divisible by 16\n if len(patches) > 0:\n # Create dataset and dataloader from patches\n patches_dataset = KidneySampleDataset(patches)\n if windows:\n patches_dataloader = DataLoader(patches_dataset, batch_size=batch_size, shuffle=False)\n else:\n patches_dataloader = DataLoader(patches_dataset, batch_size=batch_size, shuffle=False, num_workers=num_workers, prefetch_factor=prefetch_factor)\n\n # Run model\n model1.eval()\n model2.eval()\n idx = 0\n with torch.no_grad():\n for inputs, labels in patches_dataloader:\n # Clear gpu cache\n gc.collect()\n torch.cuda.empty_cache()\n # Move data to GPU\n inputs = torch.unsqueeze(torch.tensor(inputs.type(torch.cuda.FloatTensor)), axis=1)\n \n # Run and get outputs from both models\n outputs1 = model1.forward(inputs.cuda(device=0))\n outputs2 = model2.forward(inputs.cuda(device=0))\n del inputs\n\n # Transfer output tensors from gpus and convert to numpy\n out_array1 = torch.squeeze(outputs1, axis=1).cpu().detach().numpy()\n out_array2 = torch.squeeze(outputs2, axis=1).cpu().detach().numpy()\n del outputs1\n del outputs2\n \n # === Post processing model1 output ===\n \n prev_idx = idx\n # Loops through model output and transfer onto output image array\n for patch in out_array1:\n # Convert to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n\n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n \n max_y = min(idx_y+512, test_image.shape[1])\n max_x = min(idx_x+512, test_image.shape[2])\n \n out_img[idx_y: max_y, idx_x: max_x] += patch[:max_y - idx_y, :max_x - idx_x]\n idx += 1\n del out_array1\n\n # === Post processing model2 output ===\n\n idx = prev_idx\n # Loops through model output and transfer onto output image array\n for patch in out_array2:\n # Converty to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n\n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n max_y = min(idx_y+512, test_image.shape[1])\n max_x = min(idx_x+512, test_image.shape[2])\n out_img[idx_y: max_y, idx_x: max_x] = np.where(out_img[idx_y: max_y, idx_x: max_x] + patch[:max_y - idx_y, :max_x - idx_x] >= 1, 1, 0)\n idx += 1\n del out_array2\n return out_img\n\n# Generate ensemble model output from image\ndef test_stride_dataloader_v5(\n model1,\n model2,\n test_image,\n image_value_lb = 100,\n image_value_ub = 205, \n stride = 100,\n windows = False,\n batch_size = 16,\n num_workers = 8,\n prefetch_factor = 1,\n size_lowerbound = 10000,\n ):\n\n img_height = test_image.shape[1]\n img_width = test_image.shape[2]\n \n # Initialize array for output image\n out_img = np.zeros((img_height,img_width), dtype=np.uint8)\n \n \n patches = []\n patch_coordinates = []\n\n os.makedirs('../temp', exist_ok=True)\n\n num_chunk = 0\n # Segment input image into patches and storing into list\n for idx_y in tqdm(range(0, img_height, stride), desc='Segmenting Image'):\n for idx_x in range(0, img_width, stride):\n \n # Generate patch object\n center_x = idx_x + 512 // 2\n center_y = idx_y + 512 // 2\n test_patch = Patch(\n center_x = center_x,\n center_y = center_y,\n theta = 0,\n patch_size = 512,\n glomeruli = None,\n image = test_image\n )\n \n # Render patch image and test if kidney sample is present through thresholding\n # Too dark / light means the patch is empty\n image = test_patch.render_image()\n mean = np.mean(image[0])\n if mean >image_value_ub or mean = 64:\n patches_dataset = KidneySampleDataset(patches)\n patches_dataset.save(f'../temp/chunk_{num_chunk}.npy')\n np.save(f'../temp/coor_{num_chunk}.npy', np.array(patch_coordinates))\n num_chunk += 1\n\n del patches_dataset\n\n patches = []\n patch_coordinates = []\n if len(patches) > 0:\n patches_dataset = KidneySampleDataset(patches)\n patches_dataset.save(f'../temp/chunk_{num_chunk}.npy')\n np.save(f'../temp/coor_{num_chunk}.npy', np.array(patch_coordinates))\n num_chunk += 1\n\n del patches_dataset\n del patches\n del patch_coordinates\n\n del test_image\n\n for idx_chunk in tqdm(range(num_chunk), desc='Processing Chunks'):\n patches_dataset = KidneySampleDataset()\n patches_dataset.load(f'../temp/chunk_{idx_chunk}.npy')\n os.remove(f'../temp/chunk_{idx_chunk}.npy')\n\n patch_coordinates = np.load(f'../temp/coor_{idx_chunk}.npy')\n os.remove(f'../temp/coor_{idx_chunk}.npy')\n\n if windows:\n patches_dataloader = DataLoader(patches_dataset, batch_size=batch_size, shuffle=False)\n else:\n patches_dataloader = DataLoader(patches_dataset, batch_size=batch_size, shuffle=False, num_workers=num_workers, prefetch_factor=prefetch_factor)\n\n model1.eval()\n model2.eval()\n idx = 0\n with torch.no_grad():\n for inputs, labels in patches_dataloader:\n # Clear gpu cache\n gc.collect()\n torch.cuda.empty_cache()\n # Move data to GPU\n inputs = torch.unsqueeze(torch.tensor(inputs.type(torch.cuda.FloatTensor)), axis=1)\n \n # Run and get outputs from both models\n outputs1 = model1.forward(inputs.cuda(device=0))\n outputs2 = model2.forward(inputs.cuda(device=0))\n del inputs\n \n # Transfer output tensors from gpus and convert to numpy\n out_array1 = torch.squeeze(outputs1, axis=1).cpu().detach().numpy()\n out_array2 = torch.squeeze(outputs2, axis=1).cpu().detach().numpy()\n del outputs1\n del outputs2\n \n # === Post processing model1 output ===\n \n prev_idx = idx\n # Loops through model output and transfer onto output image array\n for patch in out_array1:\n # Convert to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n \n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n\n max_y = min(idx_y+512, img_height)\n max_x = min(idx_x+512, img_width)\n \n out_img[idx_y: max_y, idx_x: max_x] += patch[:max_y - idx_y, :max_x - idx_x]\n idx += 1\n del out_array1\n \n # === Post processing model2 output ===\n \n idx = prev_idx\n # Loops through model output and transfer onto output image array\n for patch in out_array2:\n # Converty to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n \n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n \n max_y = min(idx_y+512, img_height)\n max_x = min(idx_x+512, img_width)\n \n out_img[idx_y: max_y, idx_x: max_x] = np.where(out_img[idx_y: max_y, idx_x: max_x] + patch[:max_y - idx_y, :max_x - idx_x] >= 1, 1, 0)\n idx += 1\n del out_array2\n return out_img\n\ndef test_stride_dataloader_v6(\n model1,\n model2,\n test_image,\n image_value_lb = 100,\n image_value_ub = 205, \n stride = 100,\n windows = False,\n batch_size = 16,\n num_workers = 8,\n prefetch_factor = 1,\n size_lowerbound = 10000,\n ):\n \n # Initialize array for output image\n out_img = np.zeros((test_image.shape[0],test_image.shape[1]), dtype=np.uint8)\n \n \n patches = []\n patch_coordinates = []\n last_y = False\n # Segment input image into patches and storing into list\n for idx_y in tqdm(range(0, test_image.shape[0], stride), desc='Segmenting Image'):\n if idx_y + 512 > test_image.shape[0]:\n last_y = True\n idx_y = test_image.shape[0] - 512\n last_x = False\n for idx_x in range(0, test_image.shape[1], stride):\n if idx_x + 512 > test_image.shape[1]:\n last_x = True\n idx_x = test_image.shape[1] - 512\n \n patch = test_image[idx_y:idx_y+512, idx_x:idx_x+512]\n \n # Render patch image and test if kidney sample is present through thresholding\n # Too dark / light means the patch is empty\n mean = np.mean(patch)\n if mean >image_value_ub or mean = 16:\n input_list = [torch.unsqueeze(torch.from_numpy(arr), 0) for arr in patches]\n inputs = torch.stack(input_list, dim=0)\n \n \n # Run model\n model1.eval()\n model2.eval()\n idx = 0\n with torch.no_grad():\n # Clear gpu cache\n gc.collect()\n torch.cuda.empty_cache()\n # Move data to GPU\n inputs = inputs.type(torch.cuda.FloatTensor)\n\n # Run and get outputs from both models\n outputs1 = model1.forward(inputs.cuda(device=0))\n outputs2 = model2.forward(inputs.cuda(device=0))\n\n # Transfer output tensors from gpus and convert to numpy\n out_array1 = torch.squeeze(outputs1, axis=1).cpu().detach().numpy()\n out_array2 = torch.squeeze(outputs2, axis=1).cpu().detach().numpy()\n del outputs1\n del outputs2\n\n # === Post processing model1 output ===\n\n prev_idx = idx\n # Loops through model output and transfer onto output image array\n for patch in out_array1:\n # Convert to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n\n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n\n out_img[idx_y: idx_y + 512, idx_x: idx_x + 512] += patch\n idx += 1\n del out_array1\n\n # === Post processing model2 output ===\n\n idx = prev_idx\n # Loops through model output and transfer onto output image array\n for patch in out_array2:\n # Converty to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n\n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n\n out_img[idx_y: idx_y + 512, idx_x: idx_x + 512] = np.where(out_img[idx_y: idx_y + 512, idx_x: idx_x + 512] + patch >= 1, 1, 0)\n idx += 1\n del out_array2\n del patches\n del patch_coordinates\n \n patches = []\n patch_coordinates = []\n if last_x:\n break\n if last_y:\n break\n # Process last patches if number of patches isn't divisible by 16\n if len(patches) > 0:\n input_list = [torch.unsqueeze(torch.from_numpy(arr), 0) for arr in patches]\n inputs = torch.stack(input_list, dim=0)\n\n # Run model\n model1.eval()\n model2.eval()\n idx = 0\n with torch.no_grad():\n # Clear gpu cache\n gc.collect()\n torch.cuda.empty_cache()\n # Move data to GPU\n inputs = inputs.type(torch.cuda.FloatTensor)\n\n # Run and get outputs from both models\n outputs1 = model1.forward(inputs.cuda(device=0))\n outputs2 = model2.forward(inputs.cuda(device=0))\n\n # Transfer output tensors from gpus and convert to numpy\n out_array1 = torch.squeeze(outputs1, axis=1).cpu().detach().numpy()\n out_array2 = torch.squeeze(outputs2, axis=1).cpu().detach().numpy()\n del outputs1\n del outputs2\n\n # === Post processing model1 output ===\n\n prev_idx = idx\n # Loops through model output and transfer onto output image array\n for patch in out_array1:\n # Convert to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n\n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n\n out_img[idx_y: idx_y + 512, idx_x: idx_x + 512] += patch\n idx += 1\n del out_array1\n\n # === Post processing model2 output ===\n\n idx = prev_idx\n # Loops through model output and transfer onto output image array\n for patch in out_array2:\n # Converty to binary\n patch = np.where(patch >= 0.5, 1, 0).astype(np.uint8)\n\n # Find connected components and get area\n nlabels, labels, stats, centroids = cv2.connectedComponentsWithStats(patch, 8, cv2.CV_32S)\n areas = stats[1:,cv2.CC_STAT_AREA]\n\n del stats\n del centroids\n\n patch = np.zeros((labels.shape), np.uint8)\n\n # Copy over only components with areas above threshold to remove small noise\n for i in range(0, nlabels - 1):\n if areas[i] >= size_lowerbound:\n patch[labels == i + 1] = 1\n\n del nlabels\n del labels\n\n # Transfer postprocessed output to output image\n idx_x = patch_coordinates[idx][0]\n idx_y = patch_coordinates[idx][1]\n\n out_img[idx_y: idx_y + 512, idx_x: idx_x + 512] = np.where(out_img[idx_y: idx_y + 512, idx_x: idx_x + 512] + patch >= 1, 1, 0)\n idx += 1\n del out_array2\n return out_img","metadata":{"papermill":{"duration":0.187491,"end_time":"2023-11-24T00:08:17.099436","exception":false,"start_time":"2023-11-24T00:08:16.911945","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:32:23.871993Z","iopub.execute_input":"2023-11-24T14:32:23.872258Z","iopub.status.idle":"2023-11-24T14:32:24.121107Z","shell.execute_reply.started":"2023-11-24T14:32:23.872235Z","shell.execute_reply":"2023-11-24T14:32:24.120136Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def save_image_chunks(image, image_value_lb=100, image_value_ub=205):\n patches = []\n patch_coordinates = []\n\n os.mkdirs('./image_chunks')\n \n # Segment input image into patches and storing into list\n for idx_y in tqdm(range(0, image.shape[1], stride), desc='Segmenting Image'):\n for idx_x in range(0, image.shape[2], stride):\n # Generate patch object\n center_x = idx_x + 512 // 2\n center_y = idx_y + 512 // 2\n patch = Patch(\n center_x = center_x,\n center_y = center_y,\n theta = 0,\n patch_size = 512,\n glomeruli = None,\n image = image\n )\n \n # Render patch image and test if kidney sample is present through thresholding\n # Too dark / light means the patch is empty\n patch_image = patch.render_image()\n mean = np.mean(image[0])\n if mean >image_value_ub or mean = 16):\n patches_dataset = KidneySampleDataset(patches)\n \n ","metadata":{"execution":{"iopub.status.busy":"2023-11-24T14:32:24.122176Z","iopub.execute_input":"2023-11-24T14:32:24.122461Z","iopub.status.idle":"2023-11-24T14:32:24.137244Z","shell.execute_reply.started":"2023-11-24T14:32:24.122435Z","shell.execute_reply":"2023-11-24T14:32:24.136478Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import pandas as pd\ndf_sample = pd.read_csv('../input/hubmap-kidney-segmentation/sample_submission.csv')\ndisplay(df_sample)\n\nimport numba\n@numba.njit()\ndef rle_encode(pixels):\n rle_list = []\n index_pixel = 0\n count = 0\n \n for pixel in pixels:\n index_pixel += 1\n \n # Records the starting position where pixel is included\n if pixel == 1:\n if count == 0:\n start_position = index_pixel\n count += 1\n \n # Count acts as both counter of run-length and flag. \n elif count > 0:\n # records the starting position and the run-length\n rle_list.extend([start_position, count])\n # count = 0 indicates no need to run anything if current pixel = 0\n count = 0\n\n # If last pixel is still 1 --> not yet included in rle_list --> record them.\n if count > 0:\n rle_list.extend([start_position, count])\n\n return rle_list\n\ndef rle_string(binary_image):\n pixels = binary_image.flatten(order = 'F') # Top to bottom, then Left to right\n rle_list = rle_encode(pixels)\n rle_string = ' '.join(map(str, rle_list))\n return rle_string","metadata":{"papermill":{"duration":0.810295,"end_time":"2023-11-24T00:08:17.914700","exception":false,"start_time":"2023-11-24T00:08:17.104405","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:32:24.138347Z","iopub.execute_input":"2023-11-24T14:32:24.138632Z","iopub.status.idle":"2023-11-24T14:32:24.351954Z","shell.execute_reply.started":"2023-11-24T14:32:24.138607Z","shell.execute_reply":"2023-11-24T14:32:24.350967Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import time\nimport rasterio\nstart_time = time.time()\nstride = 512\ntest_images_dir = os.path.join(base_dir, 'hubmap-kidney-segmentation/test')\ntest_image_names = df_sample['id'].tolist()\n\nsubmission = {}\n\nmodel1 = UNet_Large((256, 256), (512, 512)).cuda()\nmodel1.load_state_dict(torch.load(unet_path,map_location='cuda:0'))\nmodel2 = CBAM_R2UNet_v2_Large((256, 256), (512, 512)).cuda()\nmodel2.load_state_dict(torch.load(cbam_r2unet_path,map_location='cuda:0'))\n\nfor img_id, image_name in tqdm(enumerate(test_image_names), total=len(test_image_names)):\n print(f'Processing {image_name}')\n test_image_path = os.path.join(test_images_dir, f'{image_name}.tiff')\n test_image = read_tiff(test_image_path)\n print(test_image.shape)\n # Run Models\n print('Running Model')\n out = test_stride_dataloader_v6(model1, model2, test_image, stride=stride, windows=True)\n del test_image\n gc.collect()\n\n # Plot Result\n# plt.figure(figsize=(15, 15))\n# plt.title(f'{image_name} Output (Ensemble, stride = {stride})')\n# plt.imshow(out)\n# plt.savefig(f'./test_runs/{image_name}_ensemble.png')\n \n # Encode binary image to RLE format and append to result\n print('Converting Output')\n submission[img_id] = {'id':image_name, 'predicted': rle_string(out)}\n del out\n gc.collect()\n \ndel model1\ndel model2\ngc.collect()\n \nruntime = np.round(time.time() - start_time, 2)","metadata":{"papermill":{"duration":1765.740986,"end_time":"2023-11-24T00:37:43.660661","exception":false,"start_time":"2023-11-24T00:08:17.919675","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:32:24.353371Z","iopub.execute_input":"2023-11-24T14:32:24.353695Z","iopub.status.idle":"2023-11-24T14:34:29.106176Z","shell.execute_reply.started":"2023-11-24T14:32:24.353668Z","shell.execute_reply":"2023-11-24T14:34:29.105175Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(f'Runtime = {runtime}')","metadata":{"papermill":{"duration":0.015456,"end_time":"2023-11-24T00:37:43.683710","exception":false,"start_time":"2023-11-24T00:37:43.668254","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:34:29.108933Z","iopub.execute_input":"2023-11-24T14:34:29.109230Z","iopub.status.idle":"2023-11-24T14:34:29.114372Z","shell.execute_reply.started":"2023-11-24T14:34:29.109205Z","shell.execute_reply":"2023-11-24T14:34:29.113378Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"submission = pd.DataFrame.from_dict(submission, orient='index')\nsubmission.to_csv(f'submission.csv', index=False)\nsubmission","metadata":{"papermill":{"duration":0.501335,"end_time":"2023-11-24T00:37:44.192135","exception":false,"start_time":"2023-11-24T00:37:43.690800","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:34:29.115587Z","iopub.execute_input":"2023-11-24T14:34:29.115871Z","iopub.status.idle":"2023-11-24T14:34:29.132223Z","shell.execute_reply.started":"2023-11-24T14:34:29.115846Z","shell.execute_reply":"2023-11-24T14:34:29.131083Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%reset -f out","metadata":{"papermill":{"duration":0.258396,"end_time":"2023-11-24T00:37:44.457897","exception":false,"start_time":"2023-11-24T00:37:44.199501","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2023-11-24T14:34:29.133396Z","iopub.execute_input":"2023-11-24T14:34:29.133667Z","iopub.status.idle":"2023-11-24T14:34:29.340489Z","shell.execute_reply.started":"2023-11-24T14:34:29.133644Z","shell.execute_reply":"2023-11-24T14:34:29.339340Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# df_sample = pd.read_csv('submission.csv')\n# display(df_sample)","metadata":{"execution":{"iopub.status.busy":"2023-11-24T14:34:29.341865Z","iopub.execute_input":"2023-11-24T14:34:29.342789Z","iopub.status.idle":"2023-11-24T14:34:29.350184Z","shell.execute_reply.started":"2023-11-24T14:34:29.342751Z","shell.execute_reply":"2023-11-24T14:34:29.349288Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# def rle_decode(mask_rle, shape=(256, 256)):\n# '''\n# mask_rle: run-length as string formated (start length)\n# shape: (height,width) of array to return \n# Returns numpy array, 1 - mask, 0 - background\n\n# '''\n# s = mask_rle.split()\n# starts, lengths = [np.asarray(x, dtype=int) for x in (s[0:][::2], s[1:][::2])]\n# starts -= 1\n# ends = starts + lengths\n# img = np.zeros(shape[0]*shape[1], dtype=np.uint8)\n# for lo, hi in zip(starts, ends):\n# img[lo:hi] = 1\n# return img.reshape(shape, order='F')","metadata":{"execution":{"iopub.status.busy":"2023-11-24T14:34:29.351296Z","iopub.execute_input":"2023-11-24T14:34:29.352097Z","iopub.status.idle":"2023-11-24T14:34:29.362237Z","shell.execute_reply.started":"2023-11-24T14:34:29.352064Z","shell.execute_reply":"2023-11-24T14:34:29.361482Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# plt.imshow(rle_decode(df_sample['predicted'][0], (33240, 43160)))","metadata":{"execution":{"iopub.status.busy":"2023-11-24T14:34:29.363387Z","iopub.execute_input":"2023-11-24T14:34:29.363724Z","iopub.status.idle":"2023-11-24T14:34:29.377765Z","shell.execute_reply.started":"2023-11-24T14:34:29.363698Z","shell.execute_reply":"2023-11-24T14:34:29.376647Z"},"trusted":true},"execution_count":null,"outputs":[]}]}