{"cells":[{"metadata":{"trusted":true,"_uuid":"94c3a68b408df5785a078ab8060dc1c7bccf80b3"},"cell_type":"code","source":"# necessary imports\nimport os\nimport random\nimport numpy as np\nimport pandas as pd\n\nimport torch\n\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nimport mlcrate as mlc\n\n# Add all of your code here.\n\n# Choose 56k rows for a test_1 set\n# Choose 376k rows for a test_2 set\n\n# Include a function that runs the entire model and returns best_val_f1, test_1_f1 (simulated stage 1 test f1), \n# test_2_f1 (simulated stage 2 test f1), and best_threshold\n\n# DON'T set any seeds in your code, they will be set for each run of the model in the below code cell","execution_count":null,"outputs":[]},{"metadata":{"_kg_hide-output":true,"trusted":true,"scrolled":true,"_uuid":"67ddce7eb42beb196688d8e080befd0cc870529a"},"cell_type":"code","source":"logger = mlc.LinewiseCSVWriter('shakeup.csv', header=['seed', 'val_f1', 'stage_1_test_f1', 'stage_2_test_f1', 'threshold'])\n\nfor i in range(5):\n    seed = np.random.randint(1000, high=9999)\n\n    random.seed(seed)\n    os.environ['PYTHONHASHSEED'] = str(seed)\n    np.random.seed(seed)\n    torch.manual_seed(seed)\n    torch.cuda.manual_seed(seed)\n    torch.backends.cudnn.deterministic = True\n\n    # Run cv model in function here\n    best_val_f1 = 0\n    test_1_f1 = 0\n    test_2_f1 = 0\n    best_threshold = 0\n    \n    logger.write([seed, best_val_f1, test_1_f1, test_2_f1, best_threshold])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"d59f88346b656da045444cc2987571bbd0f1892e"},"cell_type":"code","source":"log = pd.read_csv('shakeup.csv')\nlog['stage_2_shakeup'] = log['stage_2_test_f1'] - log['val_f1']\nlog['stage_1_shakeup'] = log['stage_1_test_f1'] - log['val_f1']\nlog['stage_1_2_shakeup'] = log['stage_2_test_f1'] - log['stage_1_test_f1']","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"scrolled":false,"_uuid":"d9b8a7663db0aa050433765cfd032eab7d096317"},"cell_type":"code","source":"plt.hist(log['val_f1'])\nplt.show()\nsns.kdeplot(log['val_f1'])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"6a7bec20388e01c098da2f29c1f0fbd897ab476c"},"cell_type":"code","source":"plt.hist(log['stage_1_test_f1'])\nplt.show()\nsns.kdeplot(log['stage_1_test_f1'])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"af24ceea50ab470bc567e992f9e07c133165385f"},"cell_type":"code","source":"plt.hist(log['stage_2_test_f1'])\nplt.show()\nsns.kdeplot(log['stage_2_test_f1'])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"68a40c01d088f17c92053ca4b7345491f4535432"},"cell_type":"code","source":"plt.hist(log['threshold'])\nplt.show()\nsns.kdeplot(log['threshold'])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"28a5c380330bc3967d2b7629250445bfd1b0942f"},"cell_type":"code","source":"plt.hist(log['stage_1_shakeup'])\nplt.show()\nsns.kdeplot(log['stage_1_shakeup'])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"ae3ff7b4900944854727e4021a0c4c11391ddd15"},"cell_type":"code","source":"plt.hist(log['stage_2_shakeup'])\nplt.show()\nsns.kdeplot(log['stage_2_shakeup'])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"17bac48c068e4a2064c62c0ba18118141a50edd7"},"cell_type":"code","source":"plt.hist(log['stage_1_2_shakeup'])\nplt.show()\nsns.kdeplot(log['stage_1_2_shakeup'])","execution_count":null,"outputs":[]},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","trusted":true},"cell_type":"code","source":"\n\n\n\n\n\n\n","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.6","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}