|
| | run_cnn_3class.parser = argparse.ArgumentParser(description='Run CNN over a full 2D projection.') |
| |
| | run_cnn_3class.help |
| |
| | run_cnn_3class.default |
| |
| | run_cnn_3class.args = parser.parse_args() |
| |
| | run_cnn_3class.PATCH_SIZE_W = int(args.rows) |
| |
| | run_cnn_3class.PATCH_SIZE_D = int(args.cols) |
| |
| bool | run_cnn_3class.crop_event = False |
| |
| | run_cnn_3class.rootModule = args.module |
| |
| | run_cnn_3class.rootFile = TFile(args.input) |
| |
| list | run_cnn_3class.keys = [rootModule+'/'+k.GetName()[:-4] for k in rootFile.Get(rootModule).GetListOfKeys() if '_raw' in k.GetName()] |
| |
| | run_cnn_3class.evname = keys[int(args.event)] |
| |
| | run_cnn_3class.raw |
| |
| | run_cnn_3class.deposit |
| |
| | run_cnn_3class.pdg |
| |
| | run_cnn_3class.tracks |
| |
| | run_cnn_3class.showers |
| |
| | run_cnn_3class.full2d = int(args.full) |
| |
| | run_cnn_3class.total_patches |
| |
| | run_cnn_3class.inputs = np.zeros((total_patches, PATCH_SIZE_W, PATCH_SIZE_D), dtype=np.float32) |
| |
| int | run_cnn_3class.cnt_ind = 0 |
| |
| | run_cnn_3class.model_name = args.net |
| |
| | run_cnn_3class.m = load_model(model_name) |
| |
| | run_cnn_3class.loss |
| |
| | run_cnn_3class.optimizer |
| |
| | run_cnn_3class.pred = m.predict(inputs.reshape(inputs.shape[0], 1, PATCH_SIZE_W, PATCH_SIZE_D)) |
| |
| | run_cnn_3class.outputs = np.zeros((pred.shape[1], raw.shape[0], raw.shape[1]), dtype=np.float32) |
| |
| | run_cnn_3class.mask = np.zeros(raw.shape, dtype=np.int32) |
| |
| float | run_cnn_3class.mask_thr = 0.67 |
| |
| int | run_cnn_3class.none_idx = pred.shape[1]-1 |
| |
| | run_cnn_3class.pnorm = pred[:, 0]+pred[:, 1]+pred[:, none_idx] |
| |
| float | run_cnn_3class.pn = 1.0 |
| |
| | run_cnn_3class.fig |
| |
| | run_cnn_3class.ax |
| |
| | run_cnn_3class.figsize |
| |
| | run_cnn_3class.cs = ax[0,0].pcolor(np.transpose(pdg & 0xFF), cmap='gist_ncar') |
| |
1.8.11
