{"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.12.12"},"kaggle":{"accelerator":"none","dataSources":[{"sourceType":"competition","sourceId":121150,"databundleVersionId":15833391},{"sourceType":"datasetVersion","sourceId":14899351,"datasetId":9366631,"databundleVersionId":15763950},{"sourceType":"modelInstanceVersion","sourceId":755680,"databundleVersionId":15750522,"modelInstanceId":577094,"modelId":589419}],"dockerImageVersionId":31328,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":false},"papermill":{"default_parameters":{},"duration":189.389247,"end_time":"2026-03-24T14:33:40.797412+00:00","environment_variables":{},"exception":null,"input_path":"__notebook__.ipynb","output_path":"__notebook__.ipynb","parameters":{},"start_time":"2026-03-24T14:30:31.408165+00:00","version":"2.6.0"},"widgets":{"application/vnd.jupyter.widget-state+json":{"state":{"0841860dc8924817a6509d3a6ce87034":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_89feea8626614cccbc89134d0b0a2122","placeholder":"​","style":"IPY_MODEL_aa265c6898794e56ab627d350d7addc6","tabbable":null,"tooltip":null,"value":"  [Model-B]: 100%"}},"12a6077c68d141d495f38430e7a02642":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_51bd5af13c214200b2a603fb544f7914","placeholder":"​","style":"IPY_MODEL_eb9af695d35b434a94527643d0cfcbb6","tabbable":null,"tooltip":null,"value":"Loading weights: 100%"}},"149dc5d33c7f466e8306c48030e9701c":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"161bf63c9a594288aea6185b376d148b":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"1a8399e722f04183bc1531795f32ce6d":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_5d0ce9b82654411e866e0d4b31165800","placeholder":"​","style":"IPY_MODEL_b8f1ad384b4a493197ee2ab0c340ff5c","tabbable":null,"tooltip":null,"value":" 252/252 [00:01<00:00, 238.34it/s, Materializing param=shared.weight]"}},"214f2d866d88432bb5ba30534ed109ea":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"233e31514e6d4b5eaaaf46ad895b8b3a":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"27e219f186c144aa8f5d881c15d8ffbc":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"32612b7bd1b7413d9a15a6649d9000fa":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"3945ec7013fc4d33992f02495c70692f":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"ProgressView","bar_style":"success","description":"","description_allow_html":false,"layout":"IPY_MODEL_88f214fe44aa4025830f481a9f495abe","max":4,"min":0,"orientation":"horizontal","style":"IPY_MODEL_161bf63c9a594288aea6185b376d148b","tabbable":null,"tooltip":null,"value":4}},"3caca5a78d6e489eb2203fff8940a8ba":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_233e31514e6d4b5eaaaf46ad895b8b3a","placeholder":"​","style":"IPY_MODEL_214f2d866d88432bb5ba30534ed109ea","tabbable":null,"tooltip":null,"value":" 4/4 [01:05<00:00, 15.48s/it]"}},"4a6fc240f02a401189dbe9b687132f98":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"ProgressView","bar_style":"success","description":"","description_allow_html":false,"layout":"IPY_MODEL_a18b011a9c65428aae2593a72ceab05a","max":252,"min":0,"orientation":"horizontal","style":"IPY_MODEL_809231c3e2be4bd2bf80af9369d9942a","tabbable":null,"tooltip":null,"value":252}},"4e5e09179f73466484cd4380b6b7a7a5":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_a8f3490c843a49639e01e4c3b4172d5d","placeholder":"​","style":"IPY_MODEL_9dea5d9e72fd4c79abf91c2b15bf695c","tabbable":null,"tooltip":null,"value":" 4/4 [00:00<00:00,  6.31it/s]"}},"4fe9f47411ce4de89b8e73893a9bec1e":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"ProgressView","bar_style":"success","description":"","description_allow_html":false,"layout":"IPY_MODEL_5682366ba6e5431582cfc4109e544d04","max":4,"min":0,"orientation":"horizontal","style":"IPY_MODEL_27e219f186c144aa8f5d881c15d8ffbc","tabbable":null,"tooltip":null,"value":4}},"51bd5af13c214200b2a603fb544f7914":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"5569504f29b340b38cb1fd2f7fc86866":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"ProgressView","bar_style":"success","description":"","description_allow_html":false,"layout":"IPY_MODEL_9bf3d1b722d9459ab362ac518ec856b9","max":4,"min":0,"orientation":"horizontal","style":"IPY_MODEL_6e41da56919842ce844d26a670ec9b28","tabbable":null,"tooltip":null,"value":4}},"5682366ba6e5431582cfc4109e544d04":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"5d0ce9b82654411e866e0d4b31165800":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"5eee251dcb354674870ab8c29ba5ba05":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_12a6077c68d141d495f38430e7a02642","IPY_MODEL_4a6fc240f02a401189dbe9b687132f98","IPY_MODEL_f86f7b48d0a74f7b967d16898ad1c702"],"layout":"IPY_MODEL_6c4681cb87ac4f66945fb82930b061e9","tabbable":null,"tooltip":null}},"61cb572db95543eb8fa1793e9e85e9de":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_0841860dc8924817a6509d3a6ce87034","IPY_MODEL_3945ec7013fc4d33992f02495c70692f","IPY_MODEL_3caca5a78d6e489eb2203fff8940a8ba"],"layout":"IPY_MODEL_7af9b6ab2dc04547a70d457ce316c7f6","tabbable":null,"tooltip":null}},"6acef92ecb704ebc8ebf36c2aeede3b5":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"6c4681cb87ac4f66945fb82930b061e9":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"6e41da56919842ce844d26a670ec9b28":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"71fddf9ae2ae4de1a69d720d77a2d97d":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_726c33736de045d5b896d57d0d4e3a4a","placeholder":"​","style":"IPY_MODEL_f8d3b039eae249208367ccbf957b2295","tabbable":null,"tooltip":null,"value":"  [Model-A]: 100%"}},"726c33736de045d5b896d57d0d4e3a4a":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"7af9b6ab2dc04547a70d457ce316c7f6":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"7c7544eb31a74ae2b2b66885f05cd372":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"809231c3e2be4bd2bf80af9369d9942a":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"88c4c9d5a8bc466ba78c6b355900113d":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"88f214fe44aa4025830f481a9f495abe":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"89feea8626614cccbc89134d0b0a2122":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"8c59da9cb7544286b4ba0dbd3d224657":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"ProgressView","bar_style":"success","description":"","description_allow_html":false,"layout":"IPY_MODEL_de86a1266d3c419b9fef2ee35732c5d9","max":252,"min":0,"orientation":"horizontal","style":"IPY_MODEL_88c4c9d5a8bc466ba78c6b355900113d","tabbable":null,"tooltip":null,"value":252}},"8d525a8c50ca470287cd4cb2ffe4746e":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"927c1a1ee825457181c46b1d4defb113":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"9bf3d1b722d9459ab362ac518ec856b9":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"9dea5d9e72fd4c79abf91c2b15bf695c":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"a18b011a9c65428aae2593a72ceab05a":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"a8f3490c843a49639e01e4c3b4172d5d":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"aa265c6898794e56ab627d350d7addc6":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"b4de569e208d41f3bbd3c866fab24bcd":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"b76d0a0df7b746e6b3afa2e17f436165":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"b8f1ad384b4a493197ee2ab0c340ff5c":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"baa454a4d0fc4a0e82c3c0c1b7f4972d":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"bb6df0e9d2e74e60a3196f44f7b630da":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_8d525a8c50ca470287cd4cb2ffe4746e","placeholder":"​","style":"IPY_MODEL_32612b7bd1b7413d9a15a6649d9000fa","tabbable":null,"tooltip":null,"value":"  MBR: 100%"}},"bcd4895fcd3e408594b4536995577f54":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_7c7544eb31a74ae2b2b66885f05cd372","placeholder":"​","style":"IPY_MODEL_f96a38045f0747818edc28b7ede7ec37","tabbable":null,"tooltip":null,"value":" 4/4 [00:32<00:00,  8.11s/it]"}},"c5b0931900f346559e8f471026ffbc75":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"de86a1266d3c419b9fef2ee35732c5d9":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":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,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"e1535077c4094a0cb9bf6b86d19384b4":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_efd10d9649f84ea4b7c780aef4369ca1","IPY_MODEL_8c59da9cb7544286b4ba0dbd3d224657","IPY_MODEL_1a8399e722f04183bc1531795f32ce6d"],"layout":"IPY_MODEL_b4de569e208d41f3bbd3c866fab24bcd","tabbable":null,"tooltip":null}},"e28c3b00d6b74072ac39bbbc612423b3":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_71fddf9ae2ae4de1a69d720d77a2d97d","IPY_MODEL_4fe9f47411ce4de89b8e73893a9bec1e","IPY_MODEL_bcd4895fcd3e408594b4536995577f54"],"layout":"IPY_MODEL_baa454a4d0fc4a0e82c3c0c1b7f4972d","tabbable":null,"tooltip":null}},"e7b24285bae54dc0aa829163bed574cd":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_bb6df0e9d2e74e60a3196f44f7b630da","IPY_MODEL_5569504f29b340b38cb1fd2f7fc86866","IPY_MODEL_4e5e09179f73466484cd4380b6b7a7a5"],"layout":"IPY_MODEL_149dc5d33c7f466e8306c48030e9701c","tabbable":null,"tooltip":null}},"eb9af695d35b434a94527643d0cfcbb6":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"efd10d9649f84ea4b7c780aef4369ca1":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_927c1a1ee825457181c46b1d4defb113","placeholder":"​","style":"IPY_MODEL_c5b0931900f346559e8f471026ffbc75","tabbable":null,"tooltip":null,"value":"Loading weights: 100%"}},"f86f7b48d0a74f7b967d16898ad1c702":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_6acef92ecb704ebc8ebf36c2aeede3b5","placeholder":"​","style":"IPY_MODEL_b76d0a0df7b746e6b3afa2e17f436165","tabbable":null,"tooltip":null,"value":" 252/252 [00:01<00:00, 239.75it/s, Materializing param=shared.weight]"}},"f8d3b039eae249208367ccbf957b2295":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"f96a38045f0747818edc28b7ede7ec37":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}}},"version_major":2,"version_minor":0}}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Deep Past Initiative — Akkadian-to-English Translation\n### Ensemble MBR Inference · ByT5 · Cross-model Candidate Pooling\n\nThis notebook implements a two-model ensemble with **Minimum Bayes Risk (MBR)** decoding for the [Deep Past Challenge](https://www.kaggle.com/competitions/deep-past-initiative-machine-translation).\n\n**Pipeline overview:**\n1. Preprocess Akkadian transliteration (Unicode normalisation, gap markers, fraction canonicalisation)\n2. Generate candidate translations from two ByT5 checkpoints via beam search + multi-temperature sampling\n3. Postprocess candidates (stray mark removal, shekel fractions, repeat deduplication)\n4. Select the best candidate per sample using weighted MBR (chrF++ · BLEU · Jaccard · length bonus)\n\n---","metadata":{"papermill":{"duration":0.007435,"end_time":"2026-03-24T14:30:34.013775+00:00","exception":false,"start_time":"2026-03-24T14:30:34.006340+00:00","status":"completed"},"tags":[]}},{"cell_type":"markdown","source":"## Setup & Imports","metadata":{"papermill":{"duration":0.005601,"end_time":"2026-03-24T14:30:34.025198+00:00","exception":false,"start_time":"2026-03-24T14:30:34.019597+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"import os, gc, re, json, math, random, logging, warnings\nfrom pathlib import Path\nfrom dataclasses import dataclass, field\nfrom contextlib import nullcontext\nfrom typing import List, Tuple\n\nimport pandas as pd\nimport numpy as np\nimport torch\nfrom torch.utils.data import Dataset, DataLoader, Sampler\nfrom transformers import AutoTokenizer, AutoModelForSeq2SeqLM\nfrom tqdm.auto import tqdm\nimport sacrebleu\n\nwarnings.filterwarnings(\"ignore\")\nos.environ.setdefault(\"OMP_NUM_THREADS\", \"4\")\nos.environ.setdefault(\"MKL_NUM_THREADS\", \"4\")\nos.environ.setdefault(\"TOKENIZERS_PARALLELISM\", \"true\")\n\n\n\ndef _cuda_bf16_supported() -> bool:\n if not torch.cuda.is_available():\n return False\n try:\n return bool(getattr(torch.cuda, \"is_bf16_supported\", lambda: False)())\n except Exception:\n return False\n\ndef _bf16_ctx(device: torch.device, enabled: bool):\n if enabled and device.type == \"cuda\" and _cuda_bf16_supported():\n return torch.autocast(device_type=\"cuda\", dtype=torch.bfloat16)\n return nullcontext()\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:34.038117Z","iopub.status.busy":"2026-03-24T14:30:34.037804Z","iopub.status.idle":"2026-03-24T14:30:57.515360Z","shell.execute_reply":"2026-03-24T14:30:57.514658Z"},"papermill":{"duration":23.486266,"end_time":"2026-03-24T14:30:57.517151+00:00","exception":false,"start_time":"2026-03-24T14:30:34.030885+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Configuration\n\nAll hyperparameters live in `EnsembleMBRConfig`.\nKey settings:\n\n| Parameter | Value | Description |\n|---|---|---|\n| `num_beams` | 8 | Beam width for standard beam search |\n| `sample_temperatures` | [0.55, 0.75, 0.95] | Multi-temperature sampling |\n| `mbr_w_chrf` | 0.55 | chrF++ weight in MBR scoring |\n| `mbr_w_bleu` | 0.25 | BLEU weight in MBR scoring |\n| `max_new_tokens` | 512 | Max output length (raised to prevent truncation) |\n| `mbr_pool_cap` | 48 | Max candidates fed into MBR per sample |","metadata":{"papermill":{"duration":0.006362,"end_time":"2026-03-24T14:30:57.530839+00:00","exception":false,"start_time":"2026-03-24T14:30:57.524477+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"\n@dataclass\nclass EnsembleMBRConfig:\n test_data_path: str = \"/kaggle/input/competitions/deep-past-initiative-machine-translation/test.csv\"\n output_dir: str = \"/kaggle/working/\"\n model_a_path: str = \"/kaggle/input/datasets/assiaben/final-byt5/byt5-akkadian-optimized-34x\"\n model_b_path: str = \"/kaggle/input/models/mattiaangeli/byt5-akkadian-mbr-v2/pytorch/default/1\"\n\n max_input_length: int = 512\n max_new_tokens: int = 512 \n batch_size: int = 2\n num_workers: int = 2\n num_buckets: int = 6\n\n num_beam_cands: int = 5\n num_beams: int = 8\n length_penalty: float = 1.3\n early_stopping: bool = True\n repetition_penalty: float = 1.2\n\n use_diverse_beam: bool = True\n num_diverse_cands: int = 5\n num_diverse_beams: int = 8\n num_beam_groups: int = 4\n diversity_penalty: float = 0.9\n\n use_sampling: bool = True\n sample_temperatures: List[float] = field(default_factory=lambda: [0.55, 0.75, 0.95])\n num_sample_per_temp: int = 2\n mbr_top_p: float = 0.92\n\n @property\n def num_sample_cands(self) -> int:\n return len(self.sample_temperatures) * self.num_sample_per_temp\n\n mbr_pool_cap: int = 48 \n\n mbr_w_chrf: float = 0.45 \n mbr_w_bleu: float = 0.40 \n mbr_w_jaccard: float = 0.15 \n mbr_w_length: float = 0.10\n\n use_mixed_precision: bool = False\n use_better_transformer: bool = True\n use_bucket_batching: bool = True\n use_adaptive_beams: bool = True\n aggressive_postprocessing: bool = True\n checkpoint_freq: int = 200\n\n def __post_init__(self):\n self.device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n Path(self.output_dir).mkdir(exist_ok=True, parents=True)\n\n if not torch.cuda.is_available():\n self.use_mixed_precision = False\n self.use_better_transformer = False\n\n self.use_bf16_amp = bool(\n self.use_mixed_precision\n and self.device.type == \"cuda\"\n and _cuda_bf16_supported()\n )\n\n assert self.num_beams >= self.num_beam_cands, \"num_beams must be >= num_beam_cands\"\n\n if self.use_diverse_beam:\n assert self.num_diverse_beams % self.num_beam_groups == 0\n assert self.num_diverse_beams >= self.num_diverse_cands\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.544749Z","iopub.status.busy":"2026-03-24T14:30:57.543977Z","iopub.status.idle":"2026-03-24T14:30:57.554174Z","shell.execute_reply":"2026-03-24T14:30:57.553564Z"},"papermill":{"duration":0.018628,"end_time":"2026-03-24T14:30:57.555585+00:00","exception":false,"start_time":"2026-03-24T14:30:57.536957+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Logging","metadata":{"papermill":{"duration":0.006162,"end_time":"2026-03-24T14:30:57.567933+00:00","exception":false,"start_time":"2026-03-24T14:30:57.561771+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"\ndef setup_logging(output_dir: str) -> logging.Logger:\n Path(output_dir).mkdir(exist_ok=True, parents=True)\n for h in logging.root.handlers[:]:\n logging.root.removeHandler(h)\n\n logging.basicConfig(\n level=logging.INFO,\n format=\"%(asctime)s | %(levelname)s | %(message)s\",\n handlers=[\n logging.StreamHandler(),\n logging.FileHandler(Path(output_dir) / \"ensemble_mbr.log\"),\n ],\n )\n return logging.getLogger(\"ensemble_mbr\")\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.580824Z","iopub.status.busy":"2026-03-24T14:30:57.580550Z","iopub.status.idle":"2026-03-24T14:30:57.584862Z","shell.execute_reply":"2026-03-24T14:30:57.584219Z"},"papermill":{"duration":0.012433,"end_time":"2026-03-24T14:30:57.586276+00:00","exception":false,"start_time":"2026-03-24T14:30:57.573843+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Preprocessing\n\nNormalises Akkadian transliteration before tokenisation:\n- ASCII diacritic notation → Unicode (sz→š, s,→ṣ, vowel subscripts)\n- Gap markers unified → ``\n- Determinatives standardised\n- Decimal/fraction expressions canonicalised (e.g. 0.5 → ½)\n- Whitespace collapsed","metadata":{"papermill":{"duration":0.00595,"end_time":"2026-03-24T14:30:57.598305+00:00","exception":false,"start_time":"2026-03-24T14:30:57.592355+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"\n_V2 = re.compile(r\"([aAeEiIuU])(?:2|₂)\")\n_V3 = re.compile(r\"([aAeEiIuU])(?:3|₃)\")\n_ACUTE = str.maketrans({\"a\":\"á\",\"e\":\"é\",\"i\":\"í\",\"u\":\"ú\",\"A\":\"Á\",\"E\":\"É\",\"I\":\"Í\",\"U\":\"Ú\"})\n_GRAVE = str.maketrans({\"a\":\"à\",\"e\":\"è\",\"i\":\"ì\",\"u\":\"ù\",\"A\":\"À\",\"E\":\"È\",\"I\":\"Ì\",\"U\":\"Ù\"})\n\ndef _ascii_to_diacritics(s: str) -> str:\n s = s.replace(\"sz\", \"š\").replace(\"SZ\", \"Š\")\n s = s.replace(\"s,\", \"ṣ\").replace(\"S,\", \"Ṣ\")\n s = s.replace(\"t,\", \"ṭ\").replace(\"T,\", \"Ṭ\")\n s = _V2.sub(lambda m: m.group(1).translate(_ACUTE), s)\n s = _V3.sub(lambda m: m.group(1).translate(_GRAVE), s)\n return s\n\n_ALLOWED_FRACS = [\n (1/6, \"0.16666\"), (1/4, \"0.25\"), (1/3, \"0.33333\"),\n (1/2, \"0.5\"), (2/3, \"0.66666\"), (3/4, \"0.75\"), (5/6, \"0.83333\"),\n]\n_FRAC_TOL = 2e-3\n_FLOAT_RE = re.compile(r\"(? str:\n ip = int(math.floor(x + 1e-12))\n frac = x - ip\n best = min(_ALLOWED_FRACS, key=lambda t: abs(frac - t[0]))\n if abs(frac - best[0]) <= _FRAC_TOL:\n dec = best[1]\n if ip == 0:\n return dec\n return f\"{ip}{dec[1:]}\" if dec.startswith(\"0.\") else f\"{ip}+{dec}\"\n return f\"{x:.5f}\".rstrip(\"0\").rstrip(\".\")\n\n_WS_RE = re.compile(r\"\\s+\")\n\n_GAP_UNIFIED_RE = re.compile(\n r\"<\\s*big[\\s_\\-]*gap\\s*>\"\n r\"|<\\s*gap\\s*>\"\n r\"|\\bbig[\\s_\\-]*gap\\b\"\n r\"|\\bx(?:\\s+x)+\\b\"\n r\"|\\.{3,}|…+|\\[\\.+\\]\"\n r\"|\\[\\s*x\\s*\\]|\\(\\s*x\\s*\\)\"\n r\"|(? pd.Series:\n return ser.str.replace(_GAP_UNIFIED_RE, \"\", regex=True)\n\n_CHAR_TRANS = str.maketrans({\n \"ḫ\":\"h\",\"Ḫ\":\"H\",\"ʾ\":\"\",\n \"₀\":\"0\",\"₁\":\"1\",\"₂\":\"2\",\"₃\":\"3\",\"₄\":\"4\",\n \"₅\":\"5\",\"₆\":\"6\",\"₇\":\"7\",\"₈\":\"8\",\"₉\":\"9\",\n \"—\":\"-\",\"–\":\"-\",\n})\n_SUB_X = \"ₓ\"\n\n_UNICODE_UPPER = r\"A-ZŠṬṢḪ\\u00C0-\\u00D6\\u00D8-\\u00DE\\u0160\\u1E00-\\u1EFF\"\n_UNICODE_LOWER = r\"a-zšṭṣḫ\\u00E0-\\u00F6\\u00F8-\\u00FF\\u0161\\u1E01-\\u1EFF\"\n_DET_UPPER_RE = re.compile(r\"\\(([\" + _UNICODE_UPPER + r\"0-9]{1,6})\\)\")\n_DET_LOWER_RE = re.compile(r\"\\(([\" + _UNICODE_LOWER + r\"]{1,4})\\)\")\n\n_PN_RE = re.compile(r\"\\bPN\\b\")\n_KUBABBAR_RE = re.compile(r\"KÙ\\.B\\.\")\n\n_EXACT_FRAC_RE = re.compile(r\"0\\.8333|0\\.6666|0\\.3333|0\\.1666|0\\.625|0\\.75|0\\.25|0\\.5\")\n_EXACT_FRAC_MAP = {\n \"0.8333\": \"⅚\", \"0.6666\": \"⅔\", \"0.3333\": \"⅓\", \"0.1666\": \"⅙\",\n \"0.625\": \"⅝\", \"0.75\": \"¾\", \"0.25\": \"¼\", \"0.5\": \"½\",\n}\n\ndef _frac_repl(m: re.Match) -> str:\n return _EXACT_FRAC_MAP[m.group(0)]\n\nclass OptimizedPreprocessor:\n def preprocess_batch(self, texts: List[str]) -> List[str]:\n ser = pd.Series(texts).fillna(\"\").astype(str)\n ser = ser.apply(_ascii_to_diacritics)\n ser = ser.str.replace(_DET_UPPER_RE, r\"\\1\", regex=True)\n ser = ser.str.replace(_DET_LOWER_RE, r\"{\\1}\", regex=True)\n ser = _normalize_gaps_vec(ser)\n ser = ser.str.translate(_CHAR_TRANS)\n ser = ser.str.replace(_SUB_X, \"\", regex=False)\n ser = ser.str.replace(_KUBABBAR_RE, \"KÙ.BABBAR\", regex=True)\n ser = ser.str.replace(_EXACT_FRAC_RE, _frac_repl, regex=True)\n ser = ser.str.replace(_FLOAT_RE, lambda m: _canon_decimal(float(m.group(1))), regex=True)\n ser = ser.str.replace(_WS_RE, \" \", regex=True).str.strip()\n return ser.tolist()\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.611237Z","iopub.status.busy":"2026-03-24T14:30:57.610984Z","iopub.status.idle":"2026-03-24T14:30:57.626555Z","shell.execute_reply":"2026-03-24T14:30:57.625845Z"},"papermill":{"duration":0.024074,"end_time":"2026-03-24T14:30:57.628051+00:00","exception":false,"start_time":"2026-03-24T14:30:57.603977+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Postprocessing\n\nApplied to every candidate **before** MBR scoring, so the selector compares meaningful content rather than formatting noise:\n- Gap normalisation & PN → ``\n- Shekel fraction corrections (host-confirmed: 5/12 → ⅓ shekel 15 grains)\n- Commodity regex (word-boundary safe)\n- Curly quotes → straight quotes\n- Month Roman numerals → integers\n- Stray mark removal (`<<>>`, `..`, `xx`)\n- Repeat word/phrase deduplication (phrase loop runs first)\n- Parentheses **preserved** (confirmed present in test set)","metadata":{"papermill":{"duration":0.005713,"end_time":"2026-03-24T14:30:57.639641+00:00","exception":false,"start_time":"2026-03-24T14:30:57.633928+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"\n_SOFT_GRAM_RE = re.compile(\n r\"\\(\\s*(?:fem|plur|pl|sing|singular|plural|\\?|\\!)\"\n r\"(?:\\.\\s*(?:plur|plural|sing|singular))?\"\n r\"\\.?\\s*[^)]*\\)\", re.I\n)\n_BARE_GRAM_RE = re.compile(r\"(?⌈⌋⌊[]+ʾ;')\n\n# commodity regex requires a space before the hyphen\n# so \"import-tax\" and \"kutānu-textiles\" are not affected\n_COMMODITY_RE = re.compile(r'(?<=\\s)-(gold|tax|textiles)\\b')\n_COMMODITY_REPL = {\"gold\": \"pašallum gold\", \"tax\": \"šadduātum tax\", \"textiles\": \"kutānum textiles\"}\n\ndef _commodity_repl(m: re.Match) -> str:\n return _COMMODITY_REPL[m.group(1)]\n\n# shekel fraction corrections per competition specification\n_SHEKEL_REPLS = [\n (re.compile(r'5\\s+11\\s*/\\s*12\\s+shekels?', re.I), '6 shekels less 15 grains'),\n (re.compile(r'5\\s*/\\s*12\\s+shekels?', re.I), '⅓ shekel 15 grains'),\n (re.compile(r'7\\s*/\\s*12\\s+shekels?', re.I), '½ shekel 15 grains'),\n (re.compile(r'1\\s*/\\s*12\\s*(?:\\(shekel\\)|\\bshekel)?', re.I), '15 grains'),\n]\n\n# only strip \" / or\" and \" / and\" — preserve meaningful slashes\n_SLASH_ALT_RE = re.compile(r'\\s+/\\s+(?:or|and)\\b', re.I) # FIX: no longer deletes following word\n\n_STRAY_MARKS_RE = re.compile(r'<<[^>]*>>|<(?!gap\\b)[^>]*>')\n_MULTI_GAP_RE = re.compile(r'(?:\\s*){2,}')\n\n# remove stray marks (.., xx)\n_EXTRA_STRAY_RE = re.compile(r'(? str:\n return f\"Month {_ROMAN2INT.get(m.group(1).upper(), m.group(1))}\"\n\nclass VectorizedPostprocessor:\n def postprocess_batch(self, translations: List[str]) -> List[str]:\n s = pd.Series(translations).fillna(\"\").astype(str)\n s = _normalize_gaps_vec(s)\n s = s.str.replace(_PN_RE, \"\", regex=True)\n s = s.str.replace(_COMMODITY_RE, _commodity_repl, regex=True)\n\n for pat, repl in _SHEKEL_REPLS:\n s = s.str.replace(pat, repl, regex=True)\n\n s = s.str.replace(_EXACT_FRAC_RE, _frac_repl, regex=True)\n s = s.str.replace(_FLOAT_RE, lambda m: _canon_decimal(float(m.group(1))), regex=True)\n s = s.str.replace(_SOFT_GRAM_RE, \" \", regex=True)\n s = s.str.replace(_BARE_GRAM_RE, \" \", regex=True)\n s = s.str.replace(_UNCERTAIN_RE, \"\", regex=True)\n s = s.str.replace(_STRAY_MARKS_RE, \"\", regex=True)\n s = s.str.replace(_EXTRA_STRAY_RE, \"\", regex=True) # FIX #7\n s = s.str.replace(_SLASH_ALT_RE, \"\", regex=True)\n\n # curly → straight quotes\n s = s.str.replace(_CURLY_DQ_RE, '\"', regex=True)\n s = s.str.replace(_CURLY_SQ_RE, \"'\", regex=True)\n\n s = s.str.replace(_MONTH_RE, _month_repl, regex=True)\n s = s.str.replace(_MULTI_GAP_RE, \"\", regex=True)\n\n # protect markers before stripping forbidden chars\n s = s.str.replace(\"\", \"\\x00GAP\\x00\", regex=False)\n s = s.str.translate(_FORBIDDEN_TRANS)\n s = s.str.replace(\"\\x00GAP\\x00\", \" \", regex=False)\n\n # ḫ/Ḫ → h/H\n s = s.str.translate(_HACEK_TRANS)\n # deduplicate phrases first, then single words\n for n in range(4, 1, -1):\n pat = r\"\\b((?:\\w+\\s+){\" + str(n-1) + r\"}\\w+)(?:\\s+\\1\\b)+\"\n s = s.str.replace(pat, r\"\\1\", regex=True)\n s = s.str.replace(_REPEAT_WORD_RE, r\"\\1\", regex=True) \n\n s = s.str.replace(_PUNCT_SPACE_RE, r\"\\1\", regex=True)\n s = s.str.replace(_REPEAT_PUNCT_RE, r\"\\1\", regex=True)\n s = s.str.replace(_WS_RE, \" \", regex=True).str.strip()\n return s.tolist()\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.652507Z","iopub.status.busy":"2026-03-24T14:30:57.652232Z","iopub.status.idle":"2026-03-24T14:30:57.667776Z","shell.execute_reply":"2026-03-24T14:30:57.667145Z"},"papermill":{"duration":0.023947,"end_time":"2026-03-24T14:30:57.669328+00:00","exception":false,"start_time":"2026-03-24T14:30:57.645381+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Dataset & Bucket Batching\n\n`BucketBatchSampler` groups samples by length to minimise padding waste and keep GPU utilisation stable across variable-length transliterations.","metadata":{"papermill":{"duration":0.005948,"end_time":"2026-03-24T14:30:57.681414+00:00","exception":false,"start_time":"2026-03-24T14:30:57.675466+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"\nclass AkkadianDataset(Dataset):\n def __init__(self, df: pd.DataFrame, preprocessor: OptimizedPreprocessor, logger: logging.Logger):\n self.sample_ids = df[\"id\"].tolist()\n proc = preprocessor.preprocess_batch(df[\"transliteration\"].tolist())\n self.input_texts = [\"translate Akkadian to English: \" + t for t in proc]\n logger.info(f\"Dataset: {len(self.sample_ids)} samples\")\n\n def __len__(self):\n return len(self.sample_ids)\n\n def __getitem__(self, idx):\n return self.sample_ids[idx], self.input_texts[idx]\n\nclass BucketBatchSampler(Sampler):\n def __init__(self, dataset, batch_size, num_buckets, logger, shuffle=False):\n self.batch_size = batch_size\n self.shuffle = shuffle\n lengths = [len(t.split()) for _, t in dataset]\n sorted_idx = sorted(range(len(lengths)), key=lambda i: lengths[i])\n bsize = max(1, len(sorted_idx) // max(1, num_buckets))\n self.buckets = [\n sorted_idx[i*bsize : None if i == num_buckets-1 else (i+1)*bsize]\n for i in range(num_buckets)\n ]\n\n for i, b in enumerate(self.buckets):\n if b:\n bl = [lengths[x] for x in b]\n logger.info(f\" Bucket {i}: {len(b)} samples, len [{min(bl)}, {max(bl)}]\")\n\n def __iter__(self):\n for bucket in self.buckets:\n b = list(bucket)\n if self.shuffle:\n random.shuffle(b)\n for i in range(0, len(b), self.batch_size):\n yield b[i:i+self.batch_size]\n\n def __len__(self):\n return sum(math.ceil(len(b) / self.batch_size) for b in self.buckets)\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.694705Z","iopub.status.busy":"2026-03-24T14:30:57.694045Z","iopub.status.idle":"2026-03-24T14:30:57.702825Z","shell.execute_reply":"2026-03-24T14:30:57.702164Z"},"papermill":{"duration":0.016781,"end_time":"2026-03-24T14:30:57.704247+00:00","exception":false,"start_time":"2026-03-24T14:30:57.687466+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Model Wrapper & Candidate Generation\n\nEach model produces candidates from three sources per sample:\n\n| Source | Count | Method |\n|---|---|---|\n| Standard beam | 5 | `num_beams=8`, `length_penalty=1.3` |\n| Diverse beam | 5 (if supported) | `num_beam_groups=4`, `diversity_penalty=0.9` |\n| Multi-temp sampling | 6 | temps=[0.55, 0.75, 0.95], 2 per temp |\n\nAcross two models this gives **~22 candidates** before deduplication.","metadata":{"papermill":{"duration":0.005946,"end_time":"2026-03-24T14:30:57.716284+00:00","exception":false,"start_time":"2026-03-24T14:30:57.710338+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"\nclass ModelWrapper:\n def __init__(self, model_path: str, cfg: EnsembleMBRConfig, logger: logging.Logger, label: str):\n self.cfg = cfg\n self.logger = logger\n self.label = label\n self.tokenizer = None\n self.model = None\n\n logger.info(f\"[{label}] Loading from {model_path}\")\n self.tokenizer = AutoTokenizer.from_pretrained(model_path)\n self.model = AutoModelForSeq2SeqLM.from_pretrained(model_path).to(cfg.device).eval()\n\n if cfg.device.type == \"cuda\":\n try:\n torch.set_float32_matmul_precision(\"high\")\n except Exception:\n pass\n\n n = sum(p.numel() for p in self.model.parameters())\n logger.info(f\"[{label}] {n:,} parameters\")\n\n if cfg.device.type == \"cuda\":\n used = torch.cuda.memory_allocated() / 1e9\n logger.info(f\"[{label}] GPU mem used: {used:.2f} GB\")\n\n if cfg.use_better_transformer and cfg.device.type == \"cuda\":\n try:\n from optimum.bettertransformer import BetterTransformer\n self.model = BetterTransformer.transform(self.model)\n logger.info(f\"[{label}] BetterTransformer applied\")\n except Exception as e:\n logger.warning(f\"[{label}] BetterTransformer skipped: {e}\")\n\n def collate(self, batch_samples):\n ids = [s[0] for s in batch_samples]\n texts = [s[1] for s in batch_samples]\n enc = self.tokenizer(\n texts,\n max_length=self.cfg.max_input_length,\n padding=True,\n truncation=True,\n return_tensors=\"pt\",\n )\n return ids, enc\n\n def generate_candidates(self, input_ids, attention_mask, beam_size: int) -> List[List[str]]:\n cfg = self.cfg\n B = input_ids.shape[0]\n ctx = _bf16_ctx(cfg.device, cfg.use_bf16_amp)\n\n Rb = cfg.num_beam_cands\n Rd = cfg.num_diverse_cands if cfg.use_diverse_beam else 0\n Rs = cfg.num_sample_per_temp\n\n with ctx:\n nb = max(beam_size, Rb)\n beam_out = self.model.generate(\n input_ids=input_ids,\n attention_mask=attention_mask,\n do_sample=False,\n num_beams=nb,\n num_return_sequences=Rb,\n max_new_tokens=cfg.max_new_tokens,\n length_penalty=cfg.length_penalty,\n early_stopping=cfg.early_stopping,\n repetition_penalty=cfg.repetition_penalty,\n use_cache=True,\n )\n beam_texts = self.tokenizer.batch_decode(beam_out, skip_special_tokens=True)\n\n diverse_texts = []\n actual_Rd = 0\n if cfg.use_diverse_beam:\n try:\n div_out = self.model.generate(\n input_ids=input_ids,\n attention_mask=attention_mask,\n do_sample=False,\n num_beams=cfg.num_diverse_beams,\n num_beam_groups=cfg.num_beam_groups,\n diversity_penalty=cfg.diversity_penalty,\n num_return_sequences=cfg.num_diverse_cands,\n max_new_tokens=cfg.max_new_tokens,\n length_penalty=cfg.length_penalty,\n early_stopping=cfg.early_stopping,\n repetition_penalty=cfg.repetition_penalty,\n use_cache=True,\n )\n diverse_texts = self.tokenizer.batch_decode(div_out, skip_special_tokens=True)\n actual_Rd = cfg.num_diverse_cands\n except Exception as e:\n self.logger.warning(f\"[{self.label}] Diverse beam failed ({e}), skipping\")\n\n all_samp_texts = []\n num_temps = 0\n if cfg.use_sampling and cfg.sample_temperatures:\n num_temps = len(cfg.sample_temperatures)\n for temp in cfg.sample_temperatures:\n try:\n samp_out = self.model.generate(\n input_ids=input_ids,\n attention_mask=attention_mask,\n do_sample=True,\n num_beams=1,\n top_p=cfg.mbr_top_p,\n temperature=temp,\n num_return_sequences=Rs,\n max_new_tokens=cfg.max_new_tokens,\n repetition_penalty=cfg.repetition_penalty,\n use_cache=True,\n )\n temp_texts = self.tokenizer.batch_decode(samp_out, skip_special_tokens=True)\n all_samp_texts.extend(temp_texts)\n except Exception as e:\n self.logger.warning(\n f\"[{self.label}] Sampling @ temp={temp:.2f} failed ({e}), padding with empty strings\"\n )\n all_samp_texts.extend([\"\"] * (B * Rs))\n\n pools = []\n for i in range(B):\n pool = []\n pool.extend(beam_texts[i * Rb : (i + 1) * Rb])\n\n if diverse_texts and actual_Rd > 0:\n pool.extend(diverse_texts[i * actual_Rd : (i + 1) * actual_Rd])\n\n if all_samp_texts and num_temps > 0:\n for t_idx in range(num_temps):\n seg_start = t_idx * B * Rs + i * Rs\n pool.extend(all_samp_texts[seg_start : seg_start + Rs])\n\n pools.append(pool)\n\n if pools:\n self.logger.info(\n f\"[{self.label}] Pool per sample: \"\n f\"beam={Rb} + diverse={actual_Rd} + sample={num_temps}x{Rs}={num_temps*Rs} \"\n f\"= {len(pools[0])} total\"\n )\n\n return pools\n\n def unload(self):\n label = self.label\n try:\n from optimum.bettertransformer import BetterTransformer\n self.model = BetterTransformer.reverse(self.model)\n except Exception:\n pass\n\n del self.model\n del self.tokenizer\n self.model = None\n self.tokenizer = None\n\n gc.collect()\n if torch.cuda.is_available():\n torch.cuda.empty_cache()\n torch.cuda.synchronize()\n free = (torch.cuda.get_device_properties(0).total_memory - torch.cuda.memory_allocated()) / 1e9\n self.logger.info(f\"[{label}] Unloaded. GPU free: {free:.2f} GB\")\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.729700Z","iopub.status.busy":"2026-03-24T14:30:57.729257Z","iopub.status.idle":"2026-03-24T14:30:57.746308Z","shell.execute_reply":"2026-03-24T14:30:57.745679Z"},"papermill":{"duration":0.02554,"end_time":"2026-03-24T14:30:57.747595+00:00","exception":false,"start_time":"2026-03-24T14:30:57.722055+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Weighted MBR Selector\n\nSelects the best candidate by maximising expected quality across all pairs using a weighted combination of metrics:\n\n$$\\text{score}(c_i) = \\frac{1}{N-1}\\sum_{j \\neq i} \\left[ w_{\\text{chrF}} \\cdot \\text{chrF++}(c_i, c_j) + w_{\\text{BLEU}} \\cdot \\text{BLEU}(c_i, c_j) + w_{\\text{Jaccard}} \\cdot \\text{Jaccard}(c_i, c_j) \\right] + w_{\\text{len}} \\cdot \\text{LengthBonus}(c_i)$$\n\nchrF++ is character-level and robust to OOV; BLEU captures word-level precision; Jaccard provides fast overlap regularisation; the length bonus penalises pathologically short or long outputs.","metadata":{"papermill":{"duration":0.006045,"end_time":"2026-03-24T14:30:57.759878+00:00","exception":false,"start_time":"2026-03-24T14:30:57.753833+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"\nclass MBRSelector:\n def __init__(\n self,\n pool_cap: int = 32,\n w_chrf: float = 0.55,\n w_bleu: float = 0.25,\n w_jaccard: float = 0.20,\n w_length: float = 0.10,\n ):\n self._chrf_metric = sacrebleu.metrics.CHRF(word_order=2)\n self._bleu_metric = sacrebleu.metrics.BLEU(effective_order=True)\n self.pool_cap = pool_cap\n self.w_chrf = w_chrf\n self.w_bleu = w_bleu\n self.w_jaccard = w_jaccard\n self.w_length = w_length\n self._pw_total = max(w_chrf + w_bleu + w_jaccard, 1e-9)\n\n def _chrfpp(self, a: str, b: str) -> float:\n if not a or not b:\n return 0.0\n return float(self._chrf_metric.sentence_score(a, [b]).score)\n\n def _bleu(self, a: str, b: str) -> float:\n if not a or not b:\n return 0.0\n try:\n return float(self._bleu_metric.sentence_score(a, [b]).score)\n except Exception:\n return 0.0\n\n @staticmethod\n def _jaccard(a: str, b: str) -> float:\n ta, tb = set(a.lower().split()), set(b.lower().split())\n if not ta and not tb:\n return 100.0\n if not ta or not tb:\n return 0.0\n return 100.0 * len(ta & tb) / len(ta | tb)\n\n def _pairwise_score(self, a: str, b: str) -> float:\n s = (\n self.w_chrf * self._chrfpp(a, b)\n + self.w_bleu * self._bleu(a, b)\n + self.w_jaccard * self._jaccard(a, b)\n )\n return s / self._pw_total\n\n @staticmethod\n def _length_bonus(lengths: List[int], idx: int) -> float:\n if len(lengths) == 0:\n return 100.0\n median = float(np.median(lengths))\n sigma = max(median * 0.4, 5.0)\n z = (lengths[idx] - median) / sigma\n return 100.0 * math.exp(-0.5 * z * z)\n\n @staticmethod\n def _dedup(xs: List[str]) -> List[str]:\n seen, out = set(), []\n for x in xs:\n x = str(x).strip()\n if x and x not in seen:\n out.append(x)\n seen.add(x)\n return out\n\n def pick(self, candidates: List[str]) -> str:\n cands = self._dedup(candidates)\n if self.pool_cap:\n cands = cands[:self.pool_cap]\n\n n = len(cands)\n if n == 0:\n return \"\"\n if n == 1:\n return cands[0]\n\n lengths = [len(c.split()) for c in cands]\n scores = []\n\n for i in range(n):\n pw = sum(\n self._pairwise_score(cands[i], cands[j])\n for j in range(n) if j != i\n ) / max(1, n - 1)\n\n lb = self._length_bonus(lengths, i)\n total = pw + self.w_length * lb\n scores.append(total)\n\n return cands[int(np.argmax(scores))]\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.773828Z","iopub.status.busy":"2026-03-24T14:30:57.773222Z","iopub.status.idle":"2026-03-24T14:30:57.785312Z","shell.execute_reply":"2026-03-24T14:30:57.784560Z"},"papermill":{"duration":0.020812,"end_time":"2026-03-24T14:30:57.786759+00:00","exception":false,"start_time":"2026-03-24T14:30:57.765947+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## End-to-End Engine\n\nOrchestrates the full pipeline:\n1. Load Model A → generate pools → unload\n2. Load Model B → generate pools → unload \n3. Merge pools → postprocess → MBR select → save checkpoints","metadata":{"papermill":{"duration":0.006117,"end_time":"2026-03-24T14:30:57.798984+00:00","exception":false,"start_time":"2026-03-24T14:30:57.792867+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"\nclass EnsembleMBREngine:\n def __init__(self, cfg: EnsembleMBRConfig, logger: logging.Logger):\n self.cfg = cfg\n self.logger = logger\n self.preprocessor = OptimizedPreprocessor()\n self.postprocessor = VectorizedPostprocessor()\n self.mbr = MBRSelector(\n pool_cap=cfg.mbr_pool_cap,\n w_chrf=cfg.mbr_w_chrf,\n w_bleu=cfg.mbr_w_bleu,\n w_jaccard=cfg.mbr_w_jaccard,\n w_length=cfg.mbr_w_length,\n )\n\n def _adaptive_beams(self, attn: torch.Tensor) -> int:\n if not self.cfg.use_adaptive_beams:\n return self.cfg.num_beams\n med = float(attn.sum(dim=1).float().median().item())\n short = max(self.cfg.num_beam_cands, self.cfg.num_beams // 2)\n return short if med < 100 else self.cfg.num_beams\n\n def _build_dataloader(self, dataset: AkkadianDataset, wrapper: ModelWrapper) -> DataLoader:\n if self.cfg.use_bucket_batching:\n sampler = BucketBatchSampler(dataset, self.cfg.batch_size, self.cfg.num_buckets, self.logger)\n return DataLoader(\n dataset,\n batch_sampler=sampler,\n num_workers=self.cfg.num_workers,\n collate_fn=wrapper.collate,\n pin_memory=(self.cfg.device.type == \"cuda\"),\n )\n\n return DataLoader(\n dataset,\n batch_size=self.cfg.batch_size,\n shuffle=False,\n num_workers=self.cfg.num_workers,\n collate_fn=wrapper.collate,\n pin_memory=(self.cfg.device.type == \"cuda\"),\n )\n\n def _run_one_model(self, wrapper: ModelWrapper, dataset: AkkadianDataset) -> dict:\n dl = self._build_dataloader(dataset, wrapper)\n pools_by_id = {}\n\n with torch.inference_mode():\n for batch_ids, enc in tqdm(dl, desc=f\" [{wrapper.label}]\"):\n input_ids = enc.input_ids.to(self.cfg.device, non_blocking=True)\n attn = enc.attention_mask.to(self.cfg.device, non_blocking=True)\n beam_size = self._adaptive_beams(attn)\n\n try:\n batch_pools = wrapper.generate_candidates(input_ids, attn, beam_size)\n for sid, pool in zip(batch_ids, batch_pools):\n pools_by_id[str(sid)] = pool\n except RuntimeError as e:\n if \"out of memory\" in str(e).lower():\n self.logger.error(f\"OOM in [{wrapper.label}] — skipping batch\")\n torch.cuda.empty_cache()\n for sid in batch_ids:\n pools_by_id.setdefault(str(sid), [])\n else:\n raise\n except Exception as e:\n self.logger.error(f\"[{wrapper.label}] batch error: {e}\")\n for sid in batch_ids:\n pools_by_id.setdefault(str(sid), [])\n\n if torch.cuda.is_available():\n torch.cuda.empty_cache()\n\n return pools_by_id\n\n def run(self, test_df: pd.DataFrame) -> pd.DataFrame:\n cfg, logger = self.cfg, self.logger\n\n cands_per_sample = (\n cfg.num_beam_cands\n + (cfg.num_diverse_cands if cfg.use_diverse_beam else 0)\n + cfg.num_sample_cands\n )\n\n logger.info(\"=\" * 60)\n logger.info(\"Ensemble × MBR | Cross-model candidate pooling v2\")\n logger.info(f\" Model A : {cfg.model_a_path}\")\n logger.info(f\" Model B : {cfg.model_b_path}\")\n logger.info(f\" Standard beam : {cfg.num_beam_cands} (num_beams={cfg.num_beams})\")\n logger.info(\n f\" Diverse beam : {'ON' if cfg.use_diverse_beam else 'OFF'} \"\n f\"cands={cfg.num_diverse_cands}, groups={cfg.num_beam_groups}, penalty={cfg.diversity_penalty}\"\n )\n logger.info(\n f\" Multi-temp sample : {'ON' if cfg.use_sampling else 'OFF'} \"\n f\"temps={cfg.sample_temperatures}, {cfg.num_sample_per_temp}/temp \"\n f\"→ {cfg.num_sample_cands} total\"\n )\n logger.info(\n f\" Cands/model/sample: {cands_per_sample} \"\n f\"→ pool ≈ {cands_per_sample * 2} (2 models, pre-dedup)\"\n )\n logger.info(f\" MBR pool cap : {cfg.mbr_pool_cap}\")\n logger.info(f\" BF16 AMP : {cfg.use_bf16_amp}\")\n logger.info(f\" batch_size : {cfg.batch_size}\")\n logger.info(\"=\" * 60)\n\n dataset = AkkadianDataset(test_df, self.preprocessor, logger)\n sample_ids = [str(s) for s in dataset.sample_ids]\n\n logger.info(\"Phase 1/2 — Model A inference\")\n wrapper_a = ModelWrapper(cfg.model_a_path, cfg, logger, \"Model-A\")\n pools_a = self._run_one_model(wrapper_a, dataset)\n wrapper_a.unload()\n del wrapper_a\n\n logger.info(\"Phase 2/2 — Model B inference\")\n wrapper_b = ModelWrapper(cfg.model_b_path, cfg, logger, \"Model-B\")\n pools_b = self._run_one_model(wrapper_b, dataset)\n wrapper_b.unload()\n del wrapper_b\n\n logger.info(\"Phase 3/3 — Pool merge + MBR selection\")\n results = []\n\n for sid in tqdm(sample_ids, desc=\" MBR\"):\n combined = pools_a.get(sid, []) + pools_b.get(sid, [])\n pp = self.postprocessor.postprocess_batch(combined) if combined else []\n chosen = self.mbr.pick(pp)\n\n if not chosen or not chosen.strip():\n chosen = \"\" \n\n results.append((sid, chosen))\n\n if len(results) % cfg.checkpoint_freq == 0:\n ckpt = Path(cfg.output_dir) / f\"checkpoint_{len(results)}.csv\"\n pd.DataFrame(results, columns=[\"id\", \"translation\"]).to_csv(ckpt, index=False)\n logger.info(f\" Checkpoint: {len(results)} rows → {ckpt}\")\n\n result_df = pd.DataFrame(results, columns=[\"id\", \"translation\"])\n self._validate(result_df)\n return result_df\n\n def _validate(self, df: pd.DataFrame):\n logger = self.logger\n logger.info(\"=\" * 60)\n empty = df[\"translation\"].str.strip().eq(\"\").sum()\n lens = df[\"translation\"].str.len()\n logger.info(f\"Empty : {empty} ({100*empty/max(1,len(df)):.2f}%)\")\n logger.info(f\"Len mean : {lens.mean():.1f} median: {lens.median():.1f} min: {lens.min()} max: {lens.max()}\")\n\n for idx in [0, len(df)//4, len(df)//2, 3*len(df)//4, len(df)-1]:\n row = df.iloc[idx]\n logger.info(f\" ID {row['id']}: {str(row['translation'])[:80]}\")\n\n logger.info(\"=\" * 60)\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.812567Z","iopub.status.busy":"2026-03-24T14:30:57.812222Z","iopub.status.idle":"2026-03-24T14:30:57.830955Z","shell.execute_reply":"2026-03-24T14:30:57.830376Z"},"papermill":{"duration":0.027408,"end_time":"2026-03-24T14:30:57.832429+00:00","exception":false,"start_time":"2026-03-24T14:30:57.805021+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Environment Summary","metadata":{"papermill":{"duration":0.006188,"end_time":"2026-03-24T14:30:57.845341+00:00","exception":false,"start_time":"2026-03-24T14:30:57.839153+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"\ndef print_env(cfg: EnsembleMBRConfig):\n print(f\"PyTorch : {torch.__version__}\")\n print(f\"CUDA : {torch.cuda.is_available()}\")\n if torch.cuda.is_available():\n print(f\"GPU : {torch.cuda.get_device_name(0)}\")\n mem = torch.cuda.get_device_properties(0).total_memory / 1e9\n print(f\"GPU Mem : {mem:.2f} GB\")\n print(f\"BF16 : {_cuda_bf16_supported()}\")\n print(f\"BF16 AMP : {cfg.use_bf16_amp}\")\n print()\n\n cands_per_sample = (\n cfg.num_beam_cands\n + (cfg.num_diverse_cands if cfg.use_diverse_beam else 0)\n + cfg.num_sample_cands\n )\n print(f\"Candidates/sample/model : {cands_per_sample}\")\n print(f\" ├─ standard beam : {cfg.num_beam_cands} (num_beams={cfg.num_beams})\")\n print(f\" ├─ diverse beam : {cfg.num_diverse_cands if cfg.use_diverse_beam else 0} (groups={cfg.num_beam_groups}, penalty={cfg.diversity_penalty})\")\n print(f\" └─ multi-temp sample : {cfg.num_sample_cands} ({cfg.num_sample_per_temp}/temp × {cfg.sample_temperatures})\")\n print(f\"Total pool (2 models) : ~{cands_per_sample * 2} (before dedup)\")\n print()\n","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.859689Z","iopub.status.busy":"2026-03-24T14:30:57.858975Z","iopub.status.idle":"2026-03-24T14:30:57.864529Z","shell.execute_reply":"2026-03-24T14:30:57.863924Z"},"papermill":{"duration":0.014314,"end_time":"2026-03-24T14:30:57.865938+00:00","exception":false,"start_time":"2026-03-24T14:30:57.851624+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Run Inference\n\nLoads test data, runs the full ensemble pipeline, and writes `submission.csv`.","metadata":{"papermill":{"duration":0.006176,"end_time":"2026-03-24T14:30:57.878085+00:00","exception":false,"start_time":"2026-03-24T14:30:57.871909+00:00","status":"completed"},"tags":[]}},{"cell_type":"code","source":"if __name__ == \"__main__\":\n cfg = EnsembleMBRConfig()\n logger = setup_logging(cfg.output_dir)\n\n print_env(cfg)\n\n logger.info(f\"Loading test data: {cfg.test_data_path}\")\n test_df = pd.read_csv(cfg.test_data_path, encoding=\"utf-8\")\n logger.info(f\"Test samples: {len(test_df)}\")\n\n engine = EnsembleMBREngine(cfg, logger)\n results_df = engine.run(test_df)\n\n out_path = Path(cfg.output_dir) / \"submission.csv\"\n results_df.to_csv(out_path, index=False)\n logger.info(f\"Saved → {out_path} ({len(results_df)} rows)\")\n results_df.head()\n cfg_snap = {\n \"model_a\": cfg.model_a_path,\n \"model_b\": cfg.model_b_path,\n \"num_beam_cands\": cfg.num_beam_cands,\n \"num_beams\": cfg.num_beams,\n \"length_penalty\": cfg.length_penalty,\n \"repetition_penalty\": cfg.repetition_penalty,\n \"use_diverse_beam\": cfg.use_diverse_beam,\n \"num_diverse_cands\": cfg.num_diverse_cands,\n \"num_diverse_beams\": cfg.num_diverse_beams,\n \"num_beam_groups\": cfg.num_beam_groups,\n \"diversity_penalty\": cfg.diversity_penalty,\n \"use_sampling\": cfg.use_sampling,\n \"sample_temperatures\": cfg.sample_temperatures,\n \"num_sample_per_temp\": cfg.num_sample_per_temp,\n \"num_sample_cands\": cfg.num_sample_cands,\n \"mbr_top_p\": cfg.mbr_top_p,\n \"mbr_w_chrf\": cfg.mbr_w_chrf,\n \"mbr_w_bleu\": cfg.mbr_w_bleu,\n \"mbr_w_jaccard\": cfg.mbr_w_jaccard,\n \"mbr_w_length\": cfg.mbr_w_length,\n \"mbr_pool_cap\": cfg.mbr_pool_cap,\n \"max_new_tokens\": cfg.max_new_tokens,\n \"use_bf16_amp\": cfg.use_bf16_amp,\n \"batch_size\": cfg.batch_size,\n }\n with open(Path(cfg.output_dir) / \"ensemble_mbr_config.json\", \"w\") as f:\n json.dump(cfg_snap, f, indent=2)\n\n print(\"Submission file:\", out_path)\n print(\"Config saved to:\", Path(cfg.output_dir) / \"ensemble_mbr_config.json\")","metadata":{"execution":{"iopub.execute_input":"2026-03-24T14:30:57.891637Z","iopub.status.busy":"2026-03-24T14:30:57.891322Z","iopub.status.idle":"2026-03-24T14:33:37.966694Z","shell.execute_reply":"2026-03-24T14:33:37.965866Z"},"papermill":{"duration":160.084117,"end_time":"2026-03-24T14:33:37.968167+00:00","exception":false,"start_time":"2026-03-24T14:30:57.884050+00:00","status":"completed"},"tags":[]},"outputs":[],"execution_count":null}]}