{"cells":[{"metadata":{"trusted":true},"cell_type":"code","source":"from fastai import __version__ as fastai_version","execution_count":3,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"fastai_version","execution_count":4,"outputs":[{"output_type":"execute_result","execution_count":4,"data":{"text/plain":"'1.0.51'"},"metadata":{}}]},{"metadata":{"trusted":true,"_uuid":"f02146926f9478e8a005abf19c743f1be3de39f6","scrolled":true},"cell_type":"code","source":"from fastai.tabular import *\nfrom fastai.callbacks import SaveModelCallback\nfrom fastai.metrics import mean_absolute_error,mae","execution_count":53,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"30c1f20e07a15cd91f3056208f3c8d3eb238ebf3"},"cell_type":"code","source":"path = Path(\"../input\")","execution_count":6,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"!head ../input/train.csv","execution_count":7,"outputs":[{"output_type":"stream","text":"acoustic_data,time_to_failure\r\n12,1.4690999832\r\n6,1.4690999821\r\n8,1.469099981\r\n5,1.4690999799\r\n8,1.4690999788\r\n8,1.4690999777\r\n9,1.4690999766\r\n7,1.4690999755\r\n-5,1.4690999744\r\n","name":"stdout"}]},{"metadata":{"trusted":true,"_uuid":"890d3950a6132214736c4074922521737aca80b5"},"cell_type":"code","source":"%%time\ntrain_df = pd.read_csv(path/'train.csv', dtype={'acoustic_data': np.int16, 'time_to_failure': np.float32})","execution_count":8,"outputs":[{"output_type":"stream","text":"CPU times: user 2min 23s, sys: 10.8 s, total: 2min 33s\nWall time: 2min 34s\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_df.head()","execution_count":9,"outputs":[{"output_type":"execute_result","execution_count":9,"data":{"text/plain":" acoustic_data time_to_failure\n0 12 1.4691\n1 6 1.4691\n2 8 1.4691\n3 5 1.4691\n4 8 1.4691","text/html":"
\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
acoustic_datatime_to_failure
0121.4691
161.4691
281.4691
351.4691
481.4691
\n
"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"len(train_df)","execution_count":10,"outputs":[{"output_type":"execute_result","execution_count":10,"data":{"text/plain":"629145480"},"metadata":{}}]},{"metadata":{"_uuid":"9796412b9d4fc2361ffb229e126071c24756fc2e"},"cell_type":"markdown","source":"# Preprocessing data.\nI used https://www.kaggle.com/gpreda/lanl-earthquake-eda-and-prediction"},{"metadata":{"trusted":true,"_uuid":"89261a53cabd2c2069f627ff6c4d47746c259424"},"cell_type":"code","source":"rows = 150000\nsegments = int(np.floor(train_df.shape[0] / rows))","execution_count":11,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"segments","execution_count":12,"outputs":[{"output_type":"execute_result","execution_count":12,"data":{"text/plain":"4194"},"metadata":{}}]},{"metadata":{"trusted":true,"_uuid":"5cdb7f28e08186acd27e57849c69393620954fa4"},"cell_type":"code","source":"train_X = pd.DataFrame(index=range(segments), dtype=np.float64)\ntrain_y = pd.DataFrame(index=range(segments), dtype=np.float64, columns=['time_to_failure'])","execution_count":13,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"c019e349542faf34c2993789d205cec326b1130c"},"cell_type":"code","source":"import warnings\nfrom scipy import stats\nfrom scipy.signal import hann\nimport matplotlib.pyplot as plt\nfrom scipy.signal import hilbert\nfrom scipy.signal import convolve\nfrom sklearn.svm import NuSVR, SVR\nfrom sklearn.kernel_ridge import KernelRidge\nfrom sklearn.metrics import mean_squared_error\nfrom sklearn.preprocessing import LabelEncoder\nfrom sklearn.metrics import mean_absolute_error\nfrom sklearn.preprocessing import StandardScaler\nfrom tqdm import tqdm\nfrom sklearn.linear_model import LinearRegression\nfrom sklearn.model_selection import KFold,StratifiedKFold, RepeatedKFold\nwarnings.filterwarnings(\"ignore\")","execution_count":14,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"0365dffe580dce46c07c8f361283c59076b02739"},"cell_type":"code","source":"def add_trend_feature(arr, abs_values=False):\n idx = np.array(range(len(arr)))\n if abs_values:\n arr = np.abs(arr)\n lr = LinearRegression()\n lr.fit(idx.reshape(-1, 1), arr)\n return lr.coef_[0]\n\ndef classic_sta_lta(x, length_sta, length_lta):\n sta = np.cumsum(x ** 2)\n # Convert to float\n sta = np.require(sta, dtype=np.float)\n # Copy for LTA\n lta = sta.copy()\n # Compute the STA and the LTA\n sta[length_sta:] = sta[length_sta:] - sta[:-length_sta]\n sta /= length_sta\n lta[length_lta:] = lta[length_lta:] - lta[:-length_lta]\n lta /= length_lta\n # Pad zeros\n sta[:length_lta - 1] = 0\n # Avoid division by zero by setting zero values to tiny float\n dtiny = np.finfo(0.0).tiny\n idx = lta < dtiny\n lta[idx] = dtiny\n return sta / lta","execution_count":15,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"8d8c8db2a98dac84c0046d33bb675cfa328308e9"},"cell_type":"code","source":"def create_features(seg_id, seg, X):\n xc = pd.Series(seg['acoustic_data'].values)\n zc = np.fft.fft(xc)\n \n X.loc[seg_id, 'mean'] = xc.mean()\n X.loc[seg_id, 'std'] = xc.std()\n X.loc[seg_id, 'max'] = xc.max()\n X.loc[seg_id, 'min'] = xc.min()\n \n #FFT transform values\n realFFT = np.real(zc)\n imagFFT = np.imag(zc)\n X.loc[seg_id, 'Rmean'] = realFFT.mean()\n X.loc[seg_id, 'Rstd'] = realFFT.std()\n X.loc[seg_id, 'Rmax'] = realFFT.max()\n X.loc[seg_id, 'Rmin'] = realFFT.min()\n X.loc[seg_id, 'Imean'] = imagFFT.mean()\n X.loc[seg_id, 'Istd'] = imagFFT.std()\n X.loc[seg_id, 'Imax'] = imagFFT.max()\n X.loc[seg_id, 'Imin'] = imagFFT.min()\n X.loc[seg_id, 'Rmean_last_5000'] = realFFT[-5000:].mean()\n X.loc[seg_id, 'Rstd__last_5000'] = realFFT[-5000:].std()\n X.loc[seg_id, 'Rmax_last_5000'] = realFFT[-5000:].max()\n X.loc[seg_id, 'Rmin_last_5000'] = realFFT[-5000:].min()\n X.loc[seg_id, 'Rmean_last_15000'] = realFFT[-15000:].mean()\n X.loc[seg_id, 'Rstd_last_15000'] = realFFT[-15000:].std()\n X.loc[seg_id, 'Rmax_last_15000'] = realFFT[-15000:].max()\n X.loc[seg_id, 'Rmin_last_15000'] = realFFT[-15000:].min()\n \n X.loc[seg_id, 'mean_change_abs'] = np.mean(np.diff(xc))\n X.loc[seg_id, 'mean_change_rate'] = np.mean(np.nonzero((np.diff(xc) / xc[:-1]))[0])\n X.loc[seg_id, 'abs_max'] = np.abs(xc).max()\n \n X.loc[seg_id, 'std_first_50000'] = xc[:50000].std()\n X.loc[seg_id, 'std_last_50000'] = xc[-50000:].std()\n X.loc[seg_id, 'std_first_10000'] = xc[:10000].std()\n X.loc[seg_id, 'std_last_10000'] = xc[-10000:].std()\n \n X.loc[seg_id, 'avg_first_50000'] = xc[:50000].mean()\n X.loc[seg_id, 'avg_last_50000'] = xc[-50000:].mean()\n X.loc[seg_id, 'avg_first_10000'] = xc[:10000].mean()\n X.loc[seg_id, 'avg_last_10000'] = xc[-10000:].mean()\n \n X.loc[seg_id, 'min_first_50000'] = xc[:50000].min()\n X.loc[seg_id, 'min_last_50000'] = xc[-50000:].min()\n X.loc[seg_id, 'min_first_10000'] = xc[:10000].min()\n X.loc[seg_id, 'min_last_10000'] = xc[-10000:].min()\n \n X.loc[seg_id, 'max_first_50000'] = xc[:50000].max()\n X.loc[seg_id, 'max_last_50000'] = xc[-50000:].max()\n X.loc[seg_id, 'max_first_10000'] = xc[:10000].max()\n X.loc[seg_id, 'max_last_10000'] = xc[-10000:].max()\n \n X.loc[seg_id, 'max_to_min'] = xc.max() / np.abs(xc.min())\n X.loc[seg_id, 'max_to_min_diff'] = xc.max() - np.abs(xc.min())\n X.loc[seg_id, 'sum'] = xc.sum()\n \n X.loc[seg_id, 'mean_change_rate_first_50000'] = np.mean(np.nonzero((np.diff(xc[:50000]) / xc[:50000][:-1]))[0])\n X.loc[seg_id, 'mean_change_rate_last_50000'] = np.mean(np.nonzero((np.diff(xc[-50000:]) / xc[-50000:][:-1]))[0])\n X.loc[seg_id, 'mean_change_rate_first_10000'] = np.mean(np.nonzero((np.diff(xc[:10000]) / xc[:10000][:-1]))[0])\n X.loc[seg_id, 'mean_change_rate_last_10000'] = np.mean(np.nonzero((np.diff(xc[-10000:]) / xc[-10000:][:-1]))[0])\n \n X.loc[seg_id, 'q95'] = np.quantile(xc, 0.95)\n X.loc[seg_id, 'q99'] = np.quantile(xc, 0.99)\n X.loc[seg_id, 'q05'] = np.quantile(xc, 0.05)\n X.loc[seg_id, 'q01'] = np.quantile(xc, 0.01)\n \n X.loc[seg_id, 'abs_q95'] = np.quantile(np.abs(xc), 0.95)\n X.loc[seg_id, 'abs_q99'] = np.quantile(np.abs(xc), 0.99)\n X.loc[seg_id, 'abs_q05'] = np.quantile(np.abs(xc), 0.05)\n \n X.loc[seg_id, 'trend'] = add_trend_feature(xc)\n X.loc[seg_id, 'abs_trend'] = add_trend_feature(xc, abs_values=True)\n X.loc[seg_id, 'abs_mean'] = np.abs(xc).mean()\n X.loc[seg_id, 'abs_std'] = np.abs(xc).std()\n \n X.loc[seg_id, 'mad'] = xc.mad()\n X.loc[seg_id, 'kurt'] = xc.kurtosis()\n X.loc[seg_id, 'skew'] = xc.skew()\n \n X.loc[seg_id, 'Hilbert_mean'] = np.abs(hilbert(xc)).mean()\n X.loc[seg_id, 'Hann_window_mean'] = (convolve(xc, hann(150), mode='same') / sum(hann(150))).mean()\n X.loc[seg_id, 'classic_sta_lta1_mean'] = classic_sta_lta(xc, 500, 10000).mean()\n X.loc[seg_id, 'classic_sta_lta2_mean'] = classic_sta_lta(xc, 5000, 100000).mean()\n X.loc[seg_id, 'classic_sta_lta3_mean'] = classic_sta_lta(xc, 3333, 6666).mean()\n X.loc[seg_id, 'classic_sta_lta4_mean'] = classic_sta_lta(xc, 10000, 25000).mean()\n X.loc[seg_id, 'Moving_average_700_mean'] = xc.rolling(window=700).mean().mean(skipna=True)\n X.loc[seg_id, 'Moving_average_1500_mean'] = xc.rolling(window=1500).mean().mean(skipna=True)\n X.loc[seg_id, 'Moving_average_3000_mean'] = xc.rolling(window=3000).mean().mean(skipna=True)\n X.loc[seg_id, 'Moving_average_6000_mean'] = xc.rolling(window=6000).mean().mean(skipna=True)\n ewma = pd.Series.ewm\n X.loc[seg_id, 'exp_Moving_average_300_mean'] = (ewma(xc, span=300).mean()).mean(skipna=True)\n X.loc[seg_id, 'exp_Moving_average_3000_mean'] = ewma(xc, span=3000).mean().mean(skipna=True)\n X.loc[seg_id, 'exp_Moving_average_30000_mean'] = ewma(xc, span=30000).mean().mean(skipna=True)\n no_of_std = 2\n X.loc[seg_id, 'MA_700MA_std_mean'] = xc.rolling(window=700).std().mean()\n X.loc[seg_id,'MA_700MA_BB_high_mean'] = (X.loc[seg_id, 'Moving_average_700_mean'] + no_of_std * X.loc[seg_id, 'MA_700MA_std_mean']).mean()\n X.loc[seg_id,'MA_700MA_BB_low_mean'] = (X.loc[seg_id, 'Moving_average_700_mean'] - no_of_std * X.loc[seg_id, 'MA_700MA_std_mean']).mean()\n X.loc[seg_id, 'MA_400MA_std_mean'] = xc.rolling(window=400).std().mean()\n X.loc[seg_id,'MA_400MA_BB_high_mean'] = (X.loc[seg_id, 'Moving_average_700_mean'] + no_of_std * X.loc[seg_id, 'MA_400MA_std_mean']).mean()\n X.loc[seg_id,'MA_400MA_BB_low_mean'] = (X.loc[seg_id, 'Moving_average_700_mean'] - no_of_std * X.loc[seg_id, 'MA_400MA_std_mean']).mean()\n X.loc[seg_id, 'MA_1000MA_std_mean'] = xc.rolling(window=1000).std().mean()\n \n X.loc[seg_id, 'iqr'] = np.subtract(*np.percentile(xc, [75, 25]))\n X.loc[seg_id, 'q999'] = np.quantile(xc,0.999)\n X.loc[seg_id, 'q001'] = np.quantile(xc,0.001)\n X.loc[seg_id, 'ave10'] = stats.trim_mean(xc, 0.1)\n \n for windows in [10, 100, 1000]:\n x_roll_std = xc.rolling(windows).std().dropna().values\n x_roll_mean = xc.rolling(windows).mean().dropna().values\n \n X.loc[seg_id, 'ave_roll_std_' + str(windows)] = x_roll_std.mean()\n X.loc[seg_id, 'std_roll_std_' + str(windows)] = x_roll_std.std()\n X.loc[seg_id, 'max_roll_std_' + str(windows)] = x_roll_std.max()\n X.loc[seg_id, 'min_roll_std_' + str(windows)] = x_roll_std.min()\n X.loc[seg_id, 'q01_roll_std_' + str(windows)] = np.quantile(x_roll_std, 0.01)\n X.loc[seg_id, 'q05_roll_std_' + str(windows)] = np.quantile(x_roll_std, 0.05)\n X.loc[seg_id, 'q95_roll_std_' + str(windows)] = np.quantile(x_roll_std, 0.95)\n X.loc[seg_id, 'q99_roll_std_' + str(windows)] = np.quantile(x_roll_std, 0.99)\n X.loc[seg_id, 'av_change_abs_roll_std_' + str(windows)] = np.mean(np.diff(x_roll_std))\n X.loc[seg_id, 'av_change_rate_roll_std_' + str(windows)] = np.mean(np.nonzero((np.diff(x_roll_std) / x_roll_std[:-1]))[0])\n X.loc[seg_id, 'abs_max_roll_std_' + str(windows)] = np.abs(x_roll_std).max()\n \n X.loc[seg_id, 'std_roll_mean_' + str(windows)] = x_roll_mean.std()\n X.loc[seg_id, 'max_roll_mean_' + str(windows)] = x_roll_mean.max()\n X.loc[seg_id, 'min_roll_mean_' + str(windows)] = x_roll_mean.min()\n X.loc[seg_id, 'q01_roll_mean_' + str(windows)] = np.quantile(x_roll_mean, 0.01)\n X.loc[seg_id, 'q05_roll_mean_' + str(windows)] = np.quantile(x_roll_mean, 0.05)\n X.loc[seg_id, 'q95_roll_mean_' + str(windows)] = np.quantile(x_roll_mean, 0.95)\n X.loc[seg_id, 'q99_roll_mean_' + str(windows)] = np.quantile(x_roll_mean, 0.99)\n X.loc[seg_id, 'av_change_abs_roll_mean_' + str(windows)] = np.mean(np.diff(x_roll_mean))\n X.loc[seg_id, 'av_change_rate_roll_mean_' + str(windows)] = np.mean(np.nonzero((np.diff(x_roll_mean) / x_roll_mean[:-1]))[0])\n X.loc[seg_id, 'abs_max_roll_mean_' + str(windows)] = np.abs(x_roll_mean).max()","execution_count":17,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"fc5a25685c5af8a12d4478dbe75e6a9bad0c19d1"},"cell_type":"code","source":"# iterate over all segments\nfor seg_id in tqdm(range(segments)):\n seg = train_df.iloc[seg_id*rows:seg_id*rows+rows]\n create_features(seg_id, seg, train_X)\n train_y.loc[seg_id, 'time_to_failure'] = seg['time_to_failure'].values[-1]","execution_count":18,"outputs":[{"output_type":"stream","text":"100%|██████████| 4194/4194 [23:10<00:00, 3.01it/s]\n","name":"stderr"}]},{"metadata":{"trusted":true,"_uuid":"c8b2fa785da89673e3a76b6eab4c0032789e7b43"},"cell_type":"code","source":"submission = pd.read_csv(path/'sample_submission.csv', index_col='seg_id')\ntest_X = pd.DataFrame(columns=train_X.columns, dtype=np.float64, index=submission.index)","execution_count":19,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"a391483ba9565d6e789c4052f088ea12bcf27093"},"cell_type":"code","source":"for seg_id in tqdm(test_X.index):\n seg = pd.read_csv(path/f'test/{seg_id}.csv')\n create_features(seg_id, seg, test_X)","execution_count":20,"outputs":[{"output_type":"stream","text":"100%|██████████| 2624/2624 [15:26<00:00, 2.84it/s]\n","name":"stderr"}]},{"metadata":{"trusted":true,"_uuid":"4ec1f654a82d5403d759393fa25d01f4c50625a0"},"cell_type":"code","source":"test_X.reset_index(inplace=True)","execution_count":21,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"b4d47c592cd5cc0c90b96a1710e92f51d07bdfd9"},"cell_type":"code","source":"df = pd.concat([train_X, train_y], axis=1)","execution_count":22,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"82b1d711b7b2bc5b9d1e59bd394572c94747ca9c"},"cell_type":"code","source":"df.to_feather(\"df\")\ntest_X.to_hdf('df_test.h5', key=\"df_test\")\ndf = pd.read_feather(\"df\")\ntest_df = pd.read_hdf('df_test.h5', key=\"df_test\")","execution_count":23,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"871ecc3b4af29928cceb2daa125e9c4f5d8e0034"},"cell_type":"code","source":"","execution_count":24,"outputs":[{"output_type":"execute_result","execution_count":24,"data":{"text/plain":"(4194, 148)"},"metadata":{}}]},{"metadata":{"trusted":true,"_uuid":"b214aadd4763eccdca19fc9bb03ad9a22ed458c9"},"cell_type":"code","source":"dep_var = 'time_to_failure'\ncont_names = None\ncat_names = None\n\nprocs = [FillMissing, Categorify, Normalize]\n# procs = [FillMissing, Categorify]","execution_count":25,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"6655de447bc4e83ccde7fa866d67af80349ae06e"},"cell_type":"code","source":"data = TabularDataBunch.from_df(path, df, dep_var, valid_idx=range(0,800), procs=procs, test_df=test_df)","execution_count":26,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"e5d9cabc42831249a3f8ff69f0bdee1fdb015ade"},"cell_type":"code","source":"data.show_batch(rows=10)","execution_count":27,"outputs":[{"output_type":"display_data","data":{"text/plain":"","text/html":"\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
max_first_50000max_first_10000av_change_rate_roll_mean_1000q05_roll_std_10q999std_roll_mean_10IstdMA_400MA_std_meanave_roll_std_1000std_roll_mean_100q95_roll_mean_100min_last_50000q05_roll_std_1000av_change_abs_roll_std_1000q01_roll_mean_100MA_400MA_BB_high_meanmax_roll_mean_1000RmaxMA_400MA_BB_low_meanav_change_abs_roll_std_100MA_1000MA_std_meanclassic_sta_lta1_meanq99_roll_std_100abs_maxIminabs_q05min_roll_std_10q001Rstd__last_5000av_change_rate_roll_mean_10abs_max_roll_mean_1000min_first_50000Imeansumexp_Moving_average_300_meanexp_Moving_average_3000_meanq99_roll_std_1000min_roll_mean_100Rmax_last_5000max_roll_mean_100min_roll_std_1000max_to_minMoving_average_700_meanRmean_last_15000std_roll_std_1000q99_roll_std_10mean_change_rate_last_10000mean_change_rate_last_50000av_change_abs_roll_mean_10ave_roll_std_100av_change_abs_roll_mean_1000q01_roll_std_10avg_first_10000stdave_roll_std_10max_to_min_diffImaxstd_last_50000q99_roll_mean_1000Moving_average_3000_meanav_change_abs_roll_mean_100avg_first_50000Rmin_last_15000exp_Moving_average_30000_meanq05_roll_mean_100abs_trendmean_change_rate_first_10000q95_roll_std_1000q01_roll_std_100maxabs_meanMoving_average_6000_meanav_change_rate_roll_std_10avg_last_10000q99_roll_mean_100abs_q95q01trendstd_last_10000std_first_50000q95_roll_mean_10MA_700MA_BB_high_meanminmadstd_roll_std_10std_roll_mean_1000min_roll_mean_10abs_max_roll_mean_100Rmax_last_15000meanave10Rstd_last_15000classic_sta_lta2_meanq05_roll_mean_1000q01_roll_mean_10mean_change_absabs_max_roll_mean_10q99std_first_10000abs_max_roll_std_100kurtmin_first_10000avg_last_50000mean_change_ratestd_roll_std_100q95_roll_mean_1000min_roll_mean_1000av_change_abs_roll_std_10max_last_50000q95max_roll_std_10max_roll_std_1000skewabs_stdq05_roll_std_100mean_change_rate_first_50000q95_roll_std_10Rmin_last_5000min_roll_std_100Rmeanabs_max_roll_std_1000av_change_rate_roll_mean_100max_roll_std_100min_last_10000MA_700MA_BB_low_meanq05q99_roll_mean_10RstdRmean_last_5000q01_roll_std_1000Hann_window_meanmax_roll_mean_10q01_roll_mean_1000max_last_10000classic_sta_lta4_meanMoving_average_1500_meanclassic_sta_lta3_meanMA_700MA_std_meanav_change_rate_roll_std_100q05_roll_mean_10q95_roll_std_100Hilbert_meanav_change_rate_roll_std_1000Rminiqrabs_max_roll_std_10abs_q99target
-0.2855-0.30680.7454-0.5830-0.1740-0.1954-0.1998-0.2904-0.3026-0.15420.8259-0.1039-0.5148-0.20250.3521-0.2191-0.05971.19810.36050.2203-0.30260.0482-0.1526-0.20670.19260.28140.20990.1784-0.17970.4382-0.05970.20780.00001.19811.20001.2157-0.18160.2225-0.2013-0.18210.1518-1.61321.2026-1.5201-0.1878-0.1745-2.6463-0.5759-0.7502-0.2791-0.9121-0.44291.0724-0.2003-0.2789-0.4686-0.1926-0.06180.49191.2227-1.27951.33640.35571.28041.12660.04480.5104-0.2721-0.4887-0.2772-0.08641.24220.49310.22960.0283-0.24950.2324-0.86630.5696-0.2478-0.1446-0.22830.1523-0.2776-0.2084-0.20150.1216-0.1817-0.19031.19811.2035-0.20610.66641.14990.23491.3125-0.1862-0.1628-0.2032-0.1905-0.43110.11750.85110.6815-0.19700.91620.2661-0.4137-0.0967-0.0895-0.2437-0.1638-0.9930-0.1910-0.54230.9176-0.30410.2716-0.7267-1.8252-0.16380.2138-0.1905-0.98110.36550.5219-0.1647-0.1336-0.8697-0.34701.1986-0.27600.61550.58940.77421.20920.2186-0.29740.23780.5912-0.1707-0.17840.75910.3557-0.9580-0.2437-0.15977.860896
-0.4392-0.4229-0.2846-1.6667-0.3632-0.3796-0.3760-0.6678-0.6930-0.2026-1.59470.4747-1.3124-0.2493-0.0310-0.7514-0.2924-1.42340.5821-0.0120-0.6930-0.7634-0.3230-0.42470.46570.2814-1.12760.3339-0.3085-0.1206-0.29240.3783-0.0000-1.4234-1.4248-1.4234-0.32360.2728-0.3530-0.3375-1.7019-0.5907-1.42640.3449-0.3775-0.41160.3979-0.33760.3065-0.6508-0.1500-0.9223-0.8691-0.3782-0.6773-0.1784-0.4657-0.3857-1.1058-1.43630.8947-1.38520.4523-1.4077-0.96450.01760.8294-0.8761-1.6779-0.4370-0.7023-1.4525-0.1550-1.3003-0.4927-0.71220.51330.1142-0.4150-0.3490-1.0802-0.76370.3987-0.6513-0.4171-0.13440.3744-0.3358-0.4408-1.4234-1.3877-0.4143-0.2893-1.34760.5209-0.4302-0.3924-0.5811-0.2923-0.4033-0.81600.3126-1.19250.0377-0.3934-1.5456-0.08840.1283-0.4142-0.9448-0.4426-0.3509-0.4945-0.3430-1.6414-0.6588-0.85640.4230-1.62280.5607-0.3509-0.1227-0.40330.48110.59940.5219-0.5738-0.32840.1604-1.3019-1.4235-0.4144-0.6912-0.55430.0492-1.42970.2153-0.6826-0.09510.5912-0.9196-0.6768-0.30410.4523-0.9580-0.4426-0.37830.030399311
-0.10620.4543-0.5780-0.8494-0.2897-0.3042-0.3210-0.5534-0.5765-0.1340-0.19690.3103-0.9864-0.04400.0946-0.5768-0.1145-0.40340.52810.1001-0.5765-1.3633-0.3220-0.28390.31330.2814-0.42510.3132-0.2399-0.5607-0.11450.11120.0000-0.4034-0.4037-0.4140-0.29870.2500-0.1584-0.1978-0.47440.2184-0.40500.3078-0.3070-0.39891.7153-0.8715-0.0810-0.53670.4422-0.53810.1287-0.3219-0.5703-0.0070-0.3133-0.32410.2068-0.4272-0.0022-0.86070.3531-0.5307-0.26750.0120-0.2530-0.8044-0.7810-0.2844-0.4992-0.4616-0.55181.3806-0.1801-0.48090.51331.7159-0.3842-0.1950-0.7060-0.58940.2904-0.5265-0.35760.27740.2885-0.1972-0.4214-0.4034-0.3907-0.34450.2181-0.53010.46370.6590-0.2537-0.51140.2397-0.3019-0.4929-0.28200.4427-0.6112-0.32460.20910.10920.3019-0.2938-0.5171-0.3351-0.23130.0856-0.2906-0.8579-0.5675-0.76690.3612-0.3401-0.3936-0.2313-1.0215-0.30190.46650.54220.5219-0.4715-0.2606-0.2714-1.0616-0.4033-0.2604-0.1389-0.5052-0.0778-0.4111-1.1180-0.5667-1.01400.5912-0.7892-0.5171-0.57360.3531-0.9580-0.3351-0.34706.0726995
-0.18940.10600.28821.20520.57230.76770.68380.66110.70260.2707-0.2651-2.45131.36522.1349-0.14400.63870.4189-0.3655-0.68121.20260.70261.89060.18680.9909-0.71440.28141.8481-0.55780.46350.92730.41890.23620.0000-0.3655-0.3652-0.37070.3675-0.53510.54690.71081.3288-0.7917-0.36860.80020.75820.34663.39301.39300.62350.62770.34991.3945-0.43370.68560.6472-0.31030.71441.7041-0.2934-0.3807-0.15160.0160-0.6662-0.3037-0.40030.94020.59960.60371.49921.06300.4882-0.38661.0061-0.2602-0.01730.4444-0.3998-0.95595.6894-0.03160.54160.6638-1.18070.63080.6829-0.0765-1.34850.70410.7952-0.3655-0.39370.71651.8756-0.2451-0.39530.55011.14770.39490.22031.35194.4200-0.0776-0.75500.97940.7594-0.5093-0.41430.84301.92380.33821.08461.04001.22260.69541.38470.04060.6895-0.7368-0.35690.40171.04000.89401.3519-6.2012-0.7048-0.75790.34670.5452-0.32891.1074-0.36521.2384-0.48726.61790.8665-0.37211.68100.68540.8839-0.64150.62850.55770.2881-0.66620.46741.08460.30874.374796
-0.11900.01571.03941.45070.20440.19140.21520.51870.55460.05460.0759-0.89301.4770-0.6883-0.14400.5083-0.1413-0.1602-0.5273-1.08510.55461.32730.15140.1959-0.30280.28140.9333-0.17420.15940.6666-0.14130.13960.0000-0.1602-0.1570-0.13620.2740-0.21030.32890.00040.4313-0.7827-0.1680-0.17190.21260.2462-1.89582.3218-0.66220.4920-0.98001.39450.79260.22770.5455-0.24440.30280.7203-0.1282-0.1779-1.13000.4540-0.40120.0536-0.36710.4511-1.91120.60951.27920.22040.3474-0.17770.6121-1.26480.02830.4444-0.3998-1.49292.53650.04220.66630.5300-0.29580.49340.28360.0664-0.2176-0.00060.5245-0.1602-0.16120.25983.0432-0.3126-0.3181-1.30160.12160.39490.08490.25990.5751-0.1055-0.73680.71730.2301-0.0865-0.18470.47190.65350.76580.33820.13800.28640.20151.4026-1.04330.6938-0.49860.92110.87890.13801.06400.2599-2.7357-0.5483-0.75790.28160.1516-0.90561.2985-0.16060.1241-0.19042.82762.3347-0.17310.13490.54001.0465-0.70310.58510.41451.0348-0.40120.46740.33820.21511.9008985
-0.4392-0.2165-0.0969-0.8494-0.2686-0.2864-0.3027-0.5223-0.5475-0.15780.00770.4944-0.8597-0.97020.1888-0.5070-0.13400.25170.5358-0.0071-0.5475-0.4091-0.2632-0.28050.39780.28140.20990.2613-0.24610.2995-0.13400.3442-0.00000.25170.25460.2734-0.27690.2495-0.3577-0.18950.04571.77610.25130.5819-0.2818-0.3290-0.0726-0.07510.4298-0.5067-2.0115-0.53810.8480-0.3024-0.53890.2040-0.3978-0.3667-0.06710.2444-0.49140.78740.39970.37500.2968-0.10760.5948-0.7686-0.7943-0.2808-0.37650.23750.30750.1464-0.1541-0.48090.4431-1.4393-0.3306-0.3387-0.5812-0.52070.3465-0.5011-0.3276-0.03900.3139-0.1890-0.45200.25170.2398-0.3192-0.80270.07740.42550.1144-0.2002-0.4417-0.1793-0.3149-0.51720.0989-0.48400.3020-0.29850.11180.1820-0.1149-0.4416-0.5171-0.3617-0.21740.4755-0.2743-0.8956-1.2473-0.71180.33590.51520.4017-0.21740.4402-0.31490.48110.54860.5219-0.4065-0.22560.6469-0.78550.2514-0.20100.1847-0.50520.17800.2502-0.1519-0.53560.45920.6528-0.7296-0.4388-0.09780.3997-0.9580-0.3617-0.31584.877496
-0.2151-0.34550.94620.8074-0.1215-0.0685-0.06470.12950.1233-0.07300.7918-0.10390.3767-0.05330.17000.1790-0.00790.8360-0.0796-0.45360.12330.7151-0.0999-0.1195-0.11450.2814-0.26080.1369-0.09931.3031-0.00790.15660.00000.83600.83690.8221-0.11010.0671-0.0899-0.04070.73020.43180.84070.8735-0.1059-0.04180.63561.1450-1.04960.12651.34010.3068-0.9074-0.06280.14840.15130.11450.16430.36750.8589-0.24680.27840.15750.68690.59550.2624-0.20220.14480.2688-0.10650.14120.87561.31690.32240.09340.2131-0.04861.1902-0.1416-0.03100.54160.17530.15230.1291-0.0817-0.00550.0986-0.0414-0.04100.83600.8216-0.05341.15110.5912-0.0130-0.8659-0.05250.0463-0.1986-0.1018-0.42440.24761.04560.3613-0.09460.79280.1807-0.30140.16070.3382-0.1634-0.11090.2398-0.08630.41130.61910.31420.1072-1.01541.1970-0.11091.3465-0.10180.3349-0.0788-0.33130.0584-0.05390.17050.58180.8360-0.03690.1827-0.39090.49900.85030.57760.12721.2790-0.27170.17900.13290.95020.15750.4674-0.1634-0.03482.5012972
-0.3624-0.26810.0623-0.5830-0.3212-0.3016-0.3002-0.4172-0.4473-0.20270.21220.3300-0.18920.10020.3081-0.3814-0.11090.60500.45140.7645-0.4473-0.7452-0.2992-0.36100.48790.28140.20990.3132-0.2464-0.4403-0.11090.36120.00000.60500.60450.6031-0.28220.2584-0.3620-0.23560.43961.47600.59880.5514-0.3330-0.3169-0.19540.51090.5883-0.40060.3281-0.44290.5691-0.2997-0.40020.0722-0.4879-0.26910.06650.5918-0.09730.52780.37640.62260.7947-0.0053-0.5341-0.6460-0.1002-0.3680-0.27970.5926-0.48270.6865-0.1150-0.24950.3728-0.1504-0.2921-0.2529-0.3941-0.39850.3987-0.3961-0.3471-0.33440.3624-0.2347-0.43260.60500.6387-0.3188-0.17700.77870.37790.5501-0.3375-0.3719-0.1649-0.3990-0.81490.20110.35290.6278-0.33400.24650.2736-0.1931-0.2992-0.5171-0.4207-0.3277-0.0032-0.2828-0.3323-0.8666-0.44580.36450.9878-0.3936-0.3277-0.7095-0.39900.32030.46630.5219-0.3507-0.21130.68010.07330.6048-0.35350.6175-0.22750.05510.5951-0.0591-0.4331-0.69020.5296-0.4559-0.33930.06820.37640.4674-0.4207-0.28464.644198
-0.2086-0.2165-1.0559-1.0106-0.1530-0.2445-0.2340-0.4327-0.4469-0.1595-0.29920.2446-1.3085-0.74240.1386-0.4421-0.1264-0.16570.4218-0.0048-0.4469-0.8389-0.1892-0.26040.14930.28140.20990.1473-0.2549-0.6335-0.12640.09980.0000-0.1657-0.1656-0.1616-0.12790.2295-0.3681-0.1941-0.8479-0.3144-0.1656-0.2185-0.1902-0.2250-1.1104-0.41430.3241-0.4233-0.2180-0.9223-0.3279-0.2313-0.4297-0.0993-0.1493-0.3306-0.0920-0.1595-0.0429-0.43410.1067-0.20750.0976-0.1312-1.7390-0.5628-1.3438-0.2590-0.3546-0.1552-0.6554-0.1197-0.1801-0.48090.37280.5344-0.3817-0.1387-0.6436-0.44980.2382-0.4139-0.2015-0.12230.2470-0.1935-0.1696-0.1657-0.1849-0.2340-0.9622-0.17760.3302-0.3213-0.2877-0.3719-0.1734-0.2079-0.31330.1825-0.0607-0.5183-0.2050-0.24740.11680.3610-0.2007-0.5171-0.2309-0.1723-0.2197-0.2110-1.21340.9413-0.57250.2220-0.72950.5607-0.1723-1.4413-0.20790.36410.43010.5219-0.3228-0.19740.0024-1.2758-0.1658-0.29810.0405-0.53790.2299-0.1628-0.1115-0.4406-1.45150.5912-0.5314-0.3856-1.04530.1067-0.9580-0.2309-0.25346.845799
-0.04850.57040.31250.1990-0.0374-0.0017-0.03620.07500.0764-0.06350.51910.0144-0.28910.14820.11970.1137-0.09870.6575-0.03610.28950.0764-0.64480.0295-0.1396-0.02000.2814-0.26080.04360.02080.4772-0.09870.1907-0.00000.65750.65910.6556-0.08280.09870.0460-0.0144-0.84911.27740.65691.2133-0.04320.0226-1.54441.95600.21840.07940.23350.3068-0.3669-0.03670.04740.30960.02000.08240.16830.67030.03860.1983-0.00600.58480.52910.12150.72420.41710.1220-0.12830.11260.68990.50050.44280.0478-0.0182-0.04860.47801.16320.00340.35440.11180.21950.0877-0.0688-0.17890.1546-0.0153-0.11080.65750.6663-0.02730.67780.7299-0.10831.3125-0.05930.04630.3202-0.1521-0.4121-0.15190.67710.1025-0.04300.45230.16010.77100.12780.3382-0.2269-0.04690.4955-0.05380.1927-0.26860.0680-0.0853-0.4861-0.0755-0.04690.2157-0.1521-0.7179-0.03650.09530.0863-0.04270.6019-0.69600.6575-0.04450.49391.25921.72630.65970.19630.07430.2225-0.14840.08880.09820.3114-0.00600.4674-0.22690.02773.6052961
"},"metadata":{}}]},{"metadata":{"trusted":true,"_uuid":"7256e16869575e415a4a8dd36b8157d901cf2066"},"cell_type":"code","source":"learn = tabular_learner(data,layers=[100,200,100,50], model_dir=Path(\"/tmp\"), metrics=[mae])","execution_count":54,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"e70251a46e1c3ac27cd21bcdccb2a380f091ddf6"},"cell_type":"code","source":"learn.lr_find(); learn.recorder.plot()","execution_count":55,"outputs":[{"output_type":"display_data","data":{"text/plain":"","text/html":""},"metadata":{}},{"output_type":"stream","text":"LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n","name":"stdout"},{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true,"_uuid":"a5f6ab60d06f00f23ff10662270367bcf833af68"},"cell_type":"code","source":"lr = 1e-02","execution_count":56,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"learn.fit_one_cycle(5, slice(lr))","execution_count":57,"outputs":[{"output_type":"display_data","data":{"text/plain":"","text/html":"Total time: 00:06

\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
epochtrain_lossvalid_lossmean_absolute_errortime
034.19596524.3287684.25599100:01
115.3164749.5182582.33005000:01
29.8577517.1245372.12632200:01
38.0834857.2848222.12349600:01
47.3251927.6506402.16409400:01
"},"metadata":{}}]},{"metadata":{"trusted":true,"_uuid":"681b27183682e2f427507afa00b4d203e21ab508"},"cell_type":"code","source":"learn.fit_one_cycle(5, slice(lr), callbacks=[SaveModelCallback(learn, mode='min', every='improvement', monitor='mean_absolute_error', name='best')])","execution_count":59,"outputs":[{"output_type":"display_data","data":{"text/plain":"","text/html":"Total time: 00:06

\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
epochtrain_lossvalid_lossmean_absolute_errortime
07.1280926.9348132.10085400:01
17.1024897.3017272.11574500:01
27.2820176.6187352.01961900:01
37.0900917.2032452.10524600:01
46.8241547.0394052.07060300:01
"},"metadata":{}},{"output_type":"stream","text":"Better model found at epoch 0 with mean_absolute_error value: 2.1008541584014893.\nBetter model found at epoch 2 with mean_absolute_error value: 2.0196194648742676.\n","name":"stdout"}]},{"metadata":{"trusted":true,"_uuid":"e7c0908a4c8d13b4b3bd84b0a1dbfa904c53919b"},"cell_type":"code","source":"learn.save('stage-1')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"13346f6333929353a21e2ad32e70bdfd77b946a8"},"cell_type":"code","source":"learn.load('stage-1');","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"44b76938db2044c5b04fc07f3781c04dfd258b95"},"cell_type":"code","source":"learn.unfreeze()","execution_count":60,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"f23b8b61ca5fe9b6404cbefab85ff50f6e1765a2"},"cell_type":"code","source":"learn.lr_find(); learn.recorder.plot()","execution_count":61,"outputs":[{"output_type":"display_data","data":{"text/plain":"","text/html":""},"metadata":{}},{"output_type":"stream","text":"LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n","name":"stdout"},{"output_type":"display_data","data":{"text/plain":"

","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"d564eb83b9720b21f0b760742b4b250605c849e9"},"cell_type":"code","source":"learn.fit_one_cycle(5,slice(1e-05), callbacks=[SaveModelCallback(learn, every='improvement', mode='min', monitor='mean_absolute_error', name='best')])","execution_count":62,"outputs":[{"output_type":"display_data","data":{"text/plain":"","text/html":"Total time: 00:06

\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
epochtrain_lossvalid_lossmean_absolute_errortime
06.7543896.7502032.04660100:01
16.7531886.7551732.04857800:01
26.7557076.8775912.07173200:01
36.7541786.7835632.05915400:01
46.7972086.6484202.02957400:01
"},"metadata":{}},{"output_type":"stream","text":"Better model found at epoch 0 with mean_absolute_error value: 2.0466010570526123.\nBetter model found at epoch 4 with mean_absolute_error value: 2.029573678970337.\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"!ls ~ -alh","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"learn.load('best')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"905c0162f2783bcfa41c65bd2c650fc411581d44"},"cell_type":"code","source":"preds, _ = learn.get_preds(ds_type=DatasetType.Test)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"c86df9491b846d161e261be4e8eea8dd3094e666"},"cell_type":"code","source":"subm_df = pd.DataFrame(test_df[\"seg_id\"])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"568869e012a92d9999924efd1dcc18ae0ca8e0b0"},"cell_type":"code","source":"subm_df[\"time_to_failure\"] = preds.numpy()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"fbca742892d1e1c85697cc5736f8cf4ae28b8b88"},"cell_type":"code","source":"subm_df.head()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"83c396a4704275cb04a5070955d32f60be6c7bd1"},"cell_type":"code","source":"subm_df.to_csv('submission.csv', index=False)","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}