Simulated Data Benchmark
Importing necessary libraries
[4]:
import os
import sys
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
plt.rcParams['pdf.fonttype'] = 42
plt.rcParams['font.sans-serif'] = ['Arial']
plt.rcParams['font.family'] = 'sans-serif'
BASE_DIR = r"E:\Project_Research\CauFinder_Project\CauFinder-master"
save_path = os.path.join(BASE_DIR, "benchmark", "output", "bm_simData_lin")
sys.path.append(BASE_DIR)
from CauFinder.benchmark import run_benchmark
Benchmark main function
[5]:
def benchmark_main(
noises,
causal_strengths,
save_path,
n_dataset=10,
method_for_threshold="cumulative_rate",
is_linear=False
):
#%%
evaluate_scores = []
benchmark_results = []
for i, noise in enumerate(noises):
for j, causal_strength in enumerate(causal_strengths):
print(f"noise:{noise}; causal_strength:{causal_strength}")
evaluate_res, benchmark_res = run_benchmark(
# simulation parameters
n_dataset=n_dataset,
noise_level=noise,
causal_strength=causal_strength,
is_linear=is_linear,
activation="relu",
# model parameters
n_latent=10,
n_hidden=64,
n_layers_encoder=0,
n_layers_decoder=0,
n_layers_dpd=0,
dropout_rate_encoder=0.0,
dropout_rate_decoder=0.0,
dropout_rate_dpd=0.0,
use_batch_norm='none',
use_batch_norm_dpd=True,
pdp_linear=True,
# eva parameters
threshold_method=method_for_threshold,
save_path=save_path)
evaluate_scores.append(evaluate_res)
benchmark_results.append(benchmark_res)
#%%
plt.rcParams.update({
'font.size': 12,
'pdf.fonttype': 42,
'ps.fonttype': 42,
'axes.labelsize': 14,
'axes.titlesize': 14,
'legend.fontsize': 12,
'xtick.labelsize': 12,
'ytick.labelsize': 12,
'figure.dpi': 300,
'savefig.dpi': 300,
'axes.linewidth': 1.2,
'grid.linewidth': 0.5,
'font.sans-serif': ['Arial'],
'font.family': 'sans-serif'
})
metrics_list = ['AUC', 'ACC', 'MCC', 'Precision', 'Specificity', 'Recall', 'F1_score', 'num_w_cum_rate']
# metrics_list = ['AUC', 'ACC', 'MCC', 'Specificity', 'Recall', 'F1_score', 'num_w_cum_rate']
for method in metrics_list:
fig, axs = plt.subplots(len(noises), len(causal_strengths),
figsize=(len(causal_strengths) * 5, len(noises) * 5))
# Ensure axs is always a 2D array for consistent indexing
if len(noises) == 1 and len(causal_strengths) == 1:
axs = np.array([[axs]])
elif len(noises) == 1:
axs = axs[np.newaxis, :] # Add a new axis to make it 2D
elif len(causal_strengths) == 1:
axs = axs[:, np.newaxis] # Add a new axis to make it 2D
for i, res in enumerate(evaluate_scores):
x, y = divmod(i, len(causal_strengths))
ax = axs[x, y] if len(noises) > 1 or len(causal_strengths) > 1 else axs[0]
sns.boxplot(data=res[f'{method}'].iloc[:, :n_dataset].T, ax=ax, palette='Set2')
plt.setp(ax.get_xticklabels(), rotation=45, ha='right')
ax.set_title(f'Noise: {noises[x]} | Causal Strengths: {causal_strengths[y]}', fontsize=14, pad=10)
if method != 'num_w_cum_rate':
if method == 'AUC':
ax.set_ylim(0, 1.0)
else:
ax.set_ylim(0, 1.0)
ax.grid(True)
plt.tight_layout()
fig.subplots_adjust(top=0.93)
fig.suptitle(f"Comparison of {method} Values for Different Methods", fontsize=16)
# plt.savefig(os.path.join(save_path, f'{method}_boxplot.png'), format='png')
# plt.savefig(os.path.join(save_path, f'{method}_boxplot.pdf'), format='pdf')
plt.show()
# plt.close()
# Generating radar chart for the mean values of each metric for different methods
def create_mean_radar_chart(evaluate_scores, metrics_list, save_path):
# Retrieve the list of method names
method_names = evaluate_scores[0][metrics_list[0]].index.tolist()
mean_values = {method: [] for method in method_names}
# Calculate mean values for each method
for metric in metrics_list:
if metric == "num_w_cum_rate":
continue # Skip num_w_cum_rate
print(f"Processing metric: {metric}") # Debug: Print current metric
for method_name in method_names:
all_values = []
for res in evaluate_scores:
value = res[metric].loc[method_name, 'Mean']
all_values.append(value)
mean_val = np.mean(all_values)
mean_values[method_name].append(mean_val)
# Define the categories excluding "num_w_cum_rate"
categories = [metric for metric in metrics_list if metric != "num_w_cum_rate"]
num_categories = len(categories)
# Calculate angles for each category
angles = [n / float(num_categories) * 2 * np.pi for n in range(num_categories)]
angles += angles[:1]
# Create radar chart
fig, ax = plt.subplots(figsize=(8, 8), subplot_kw=dict(polar=True))
# Plot and fill for each method
for method_name in method_names:
values = mean_values[method_name]
print(
f"{method_name} values length: {len(values)}, num_categories: {num_categories}") # Debug: Print lengths
if len(values) != num_categories:
print(f"Skipping {method_name} due to mismatch in number of categories.")
continue
values += values[:1]
# Debug: Print values before plotting
print(f"Plotting {method_name}: {values}")
ax.plot(angles, values, linewidth=2, linestyle='solid', label=method_name)
ax.fill(angles, values, alpha=0.25)
# Remove y-axis labels
# ax.set_yticklabels([])
# Add numbers to the circles
ax.set_rlabel_position(0)
ax.yaxis.set_tick_params(labelsize=10)
# Set x-ticks and labels
ax.set_xticks(angles[:-1])
ax.set_xticklabels(categories, fontsize=12)
# Configure legend
ax.legend(loc='upper right', bbox_to_anchor=(1.3, 1.1), fontsize=12)
# Save the chart as PDF and PNG
# plt.savefig(os.path.join(save_path, "mean_radar_chart.pdf"), format="pdf", bbox_inches='tight')
# plt.savefig(os.path.join(save_path, "mean_radar_chart.png"), format="png", bbox_inches='tight')
plt.show()
# plt.close()
create_mean_radar_chart(evaluate_scores, metrics_list, save_path)
return evaluate_scores, benchmark_results
Run benchmark
[6]:
if __name__ == '__main__':
# noises = [0.05, 0.1, 1.0]
noises = [0.01, 0.1, 1.0]
# causal_strengths = [3, 5, 8]
# causal_strengths = [2, 5, 8]
causal_strengths = [2, 4, 6]
scores, results = benchmark_main(
noises=noises,
causal_strengths=causal_strengths,
save_path=save_path,
n_dataset=10,
is_linear=True,
method_for_threshold="top_k"
)
The dtype argument is deprecated and will be removed in late 2024.
noise:0.01; causal_strength:2
This is the 1/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.86it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 2/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.85it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 3/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.74it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 4/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:28<00:00, 10.56it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 5/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:28<00:00, 10.69it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 6/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 11.01it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 7/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:28<00:00, 10.58it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 8/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:28<00:00, 10.60it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 9/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:39<00:00, 7.64it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 10/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 8.95it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
The dtype argument is deprecated and will be removed in late 2024.
noise:0.01; causal_strength:4
This is the 1/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:32<00:00, 9.30it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 2/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.56it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 3/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 8.97it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 4/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 9.06it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 5/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:32<00:00, 9.35it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 6/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 9.02it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 7/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.77it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 8/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.39it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 9/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:34<00:00, 8.73it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 10/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 8.92it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
The dtype argument is deprecated and will be removed in late 2024.
noise:0.01; causal_strength:6
This is the 1/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 8.97it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 2/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:36<00:00, 8.22it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 3/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:35<00:00, 8.51it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 4/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:29<00:00, 10.14it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 5/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:34<00:00, 8.82it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 6/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:32<00:00, 9.37it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 7/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:32<00:00, 9.33it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 8/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.75it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 9/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.55it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 10/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.95it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
The dtype argument is deprecated and will be removed in late 2024.
noise:0.1; causal_strength:2
This is the 1/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:29<00:00, 10.01it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 2/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 8.90it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 3/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.92it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 4/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:28<00:00, 10.45it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 5/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:29<00:00, 10.18it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 6/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:32<00:00, 9.28it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 7/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.72it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 8/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.90it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 9/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.52it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 10/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.93it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
The dtype argument is deprecated and will be removed in late 2024.
noise:0.1; causal_strength:4
This is the 1/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.82it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 2/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.51it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 3/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.44it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 4/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.68it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 5/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:32<00:00, 9.17it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 6/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.38it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 7/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.90it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 8/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.59it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 9/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 8.92it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 10/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:31<00:00, 9.53it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
The dtype argument is deprecated and will be removed in late 2024.
noise:0.1; causal_strength:6
This is the 1/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:32<00:00, 9.15it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 2/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 8.86it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 3/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:34<00:00, 8.69it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 4/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 8.91it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 5/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:32<00:00, 9.17it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 6/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:33<00:00, 9.04it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 7/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.86it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 8/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.77it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 9/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.32it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 10/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.31it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
The dtype argument is deprecated and will be removed in late 2024.
noise:1.0; causal_strength:2
This is the 1/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.30it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 2/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.77it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 3/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:28<00:00, 10.66it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 4/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.94it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 5/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.98it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 6/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.90it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 7/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.83it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 8/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.73it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 9/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:28<00:00, 10.63it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 10/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.33it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
The dtype argument is deprecated and will be removed in late 2024.
noise:1.0; causal_strength:4
This is the 1/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.45it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 2/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.94it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 3/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.37it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 4/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 10.87it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 5/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.14it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 6/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:25<00:00, 11.63it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 7/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.29it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 8/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:30<00:00, 9.91it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 9/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.43it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 10/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.36it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
The dtype argument is deprecated and will be removed in late 2024.
noise:1.0; causal_strength:6
This is the 1/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:25<00:00, 11.86it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
This is the 2/10 dataset
The dtype argument is deprecated and will be removed in late 2024.
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:25<00:00, 11.97it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 3/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.39it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 4/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:25<00:00, 11.71it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 5/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.36it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 6/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:25<00:00, 11.76it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 7/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:25<00:00, 11.74it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 8/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:25<00:00, 11.54it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 9/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:27<00:00, 11.10it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
The dtype argument is deprecated and will be removed in late 2024.
This is the 10/10 dataset
training: 100%|██████████████████████████████████████████████████████████████████████| 300/300 [00:26<00:00, 11.15it/s]
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
WARNING:fontTools.subset:meta NOT subset; don't know how to subset; dropped
Processing metric: AUC
Processing metric: ACC
Processing metric: MCC
Processing metric: Precision
Processing metric: Specificity
Processing metric: Recall
Processing metric: F1_score
CauFinder values length: 7, num_categories: 7
Plotting CauFinder: [0.7969506172839506, 0.9091111111111112, 0.49506172839506174, 0.5455555555555556, 0.9495061728395063, 0.5455555555555556, 0.5455555555555556, 0.7969506172839506]
PC values length: 7, num_categories: 7
Plotting PC: [0.6957037037037037, 0.8784444444444444, 0.32469135802469135, 0.3922222222222222, 0.9324691358024693, 0.3922222222222222, 0.39222222222222225, 0.6957037037037037]
VAE_SHAP values length: 7, num_categories: 7
Plotting VAE_SHAP: [0.6414197530864197, 0.8604444444444446, 0.2246913580246914, 0.3022222222222222, 0.9224691358024693, 0.3022222222222222, 0.3022222222222222, 0.6414197530864197]
VAE_Grad values length: 7, num_categories: 7
Plotting VAE_Grad: [0.7033703703703704, 0.8737777777777778, 0.29876543209876544, 0.36888888888888893, 0.9298765432098767, 0.36888888888888893, 0.36888888888888893, 0.7033703703703704]
RF values length: 7, num_categories: 7
Plotting RF: [0.5509691358024691, 0.8380000000000001, 0.09999999999999999, 0.19, 0.9099999999999999, 0.19, 0.19000000000000003, 0.5509691358024691]
MI values length: 7, num_categories: 7
Plotting MI: [0.514216049382716, 0.832888888888889, 0.07160493827160493, 0.16444444444444445, 0.9071604938271606, 0.16444444444444445, 0.16444444444444445, 0.514216049382716]
T-test values length: 7, num_categories: 7
Plotting T-test: [0.5127592592592592, 0.8346666666666667, 0.08148148148148147, 0.1733333333333333, 0.9081481481481481, 0.1733333333333333, 0.1733333333333333, 0.5127592592592592]
PCC values length: 7, num_categories: 7
Plotting PCC: [0.5129753086419754, 0.8348888888888888, 0.08271604938271604, 0.17444444444444443, 0.9082716049382716, 0.17444444444444443, 0.17444444444444443, 0.5129753086419754]
SCC values length: 7, num_categories: 7
Plotting SCC: [0.5115308641975308, 0.8355555555555556, 0.08641975308641976, 0.17777777777777776, 0.9086419753086419, 0.17777777777777776, 0.17777777777777776, 0.5115308641975308]
