def plot_new_samples(run_id, model_dir, trained):
dump_dir = "/home/akajal/WatChMaL/VAE/dumps/" + run_id + "/"
model_status = "trained" if trained is True else "untrained"
np_arr_path = dump_dir + "samples/" + model_dir + "_" + model_status + ".npz"
np_arr = np.load(np_arr_path)
np_samples, np_labels, np_energies = np_arr["samples"], np_arr["predicted_labels"], np_arr["predicted_energies"]
np_samples = np_samples.reshape(-1, np_samples.shape[2], np_samples.shape[3], np_samples.shape[4])
np_labels = np_labels.reshape(-1, 1)
np_energies = np_energies.reshape(-1, 1)
i, j = random.randint(0, np_labels.shape[0]-1), random.randint(0, np_labels.shape[0]-1)
plot_utils.plot_actual_vs_recon(np_samples[i], np_samples[j],
label_dict[np_labels[i].item()], np_energies[i].item(),
label_dict[np_labels[j].item()], np_energies[j].item(),
show_plot=True)
plot_utils.plot_charge_hist(np_samples[i],
np_samples[j], 0, num_bins=200)
plot_utils.plot_charge_hist(np_samples,
np_samples, 0, num_bins=200)
def plot_old_events(run_id, iteration, mode):
dump_dir = "/home/akajal/WatChMaL/VAE/dumps/" + run_id + "/"
if mode is "validation":
np_arr_path = dump_dir + "val_iteration_" + str(iteration) + ".npz"
else:
np_arr_path = dump_dir + "iteration_" + str(iteration) + ".npz"
# Load the numpy array
np_arr = np.load(np_arr_path)
np_event, np_recon, np_labels, np_energies = np_arr["events"], np_arr[
"recon"], np_arr["labels"], np_arr["energies"]
i = random.randint(0, np_labels.shape[0] - 1)
plot_utils.plot_actual_vs_recon(np_event[i],
np_recon[i],
label_dict[np_labels[i]],
np_energies[i].item(),
label_dict[np_labels[i]],
np_energies[i].item(),
show_plot=True)
plot_utils.plot_charge_hist(torch.tensor(np_event).permute(0, 2, 3,
1).numpy(),
np_recon,
iteration,
num_bins=200)
def plot_samples(run_id, model_dir, trained):
dump_dir = "/home/akajal/WatChMaL/VAE/dumps/" + run_id + "/"
model_status = "trained" if trained is True else "untrained"
np_arr_path = dump_dir + "samples/" + model_dir + "_" + model_status + "_samples.npy"
np_arr = np.load(np_arr_path, allow_pickle=True)
i, j = random.randint(0, np_arr.shape[0]-1), random.randint(0, np_arr.shape[0]-1)
plot_utils.plot_actual_vs_recon(np_arr[i][0][0], np_arr[j][0][0],
label_dict[np_arr[i][1].item()], np_arr[i][2][0],
show_plot=True)
plot_utils.plot_charge_hist(np_arr[i][0][0],
np_arr[j][0][0], 0, num_bins=200)