Spaces:
Sleeping
Sleeping
File size: 11,346 Bytes
ff0e79e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 | """
Plotting utilities for training metrics visualization
"""
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
from typing import Dict, List, Optional
from pathlib import Path
import json
def set_style():
"""Set matplotlib style"""
plt.style.use('seaborn-v0_8-whitegrid')
sns.set_palette("husl")
def plot_training_curves(history: Dict,
save_path: str,
title: str = "Training Progress"):
"""
Plot training and validation curves
Args:
history: Training history dictionary
save_path: Path to save plot
title: Plot title
"""
set_style()
fig, axes = plt.subplots(2, 3, figsize=(15, 10))
fig.suptitle(title, fontsize=14, fontweight='bold')
epochs = range(1, len(history.get('train_loss', [])) + 1)
# Loss
ax = axes[0, 0]
if 'train_loss' in history and history['train_loss']:
ax.plot(epochs, history['train_loss'], 'b-', label='Train', linewidth=2)
if 'val_loss' in history and history['val_loss']:
ax.plot(epochs, history['val_loss'], 'r-', label='Val', linewidth=2)
ax.set_xlabel('Epoch')
ax.set_ylabel('Loss')
ax.set_title('Loss')
ax.legend()
ax.grid(True, alpha=0.3)
# IoU
ax = axes[0, 1]
if 'train_iou' in history and history['train_iou']:
ax.plot(epochs, history['train_iou'], 'b-', label='Train', linewidth=2)
if 'val_iou' in history and history['val_iou']:
ax.plot(epochs, history['val_iou'], 'r-', label='Val', linewidth=2)
ax.set_xlabel('Epoch')
ax.set_ylabel('IoU')
ax.set_title('Intersection over Union')
ax.legend()
ax.grid(True, alpha=0.3)
# Dice
ax = axes[0, 2]
if 'train_dice' in history and history['train_dice']:
ax.plot(epochs, history['train_dice'], 'b-', label='Train', linewidth=2)
if 'val_dice' in history and history['val_dice']:
ax.plot(epochs, history['val_dice'], 'r-', label='Val', linewidth=2)
ax.set_xlabel('Epoch')
ax.set_ylabel('Dice')
ax.set_title('Dice Score (F1)')
ax.legend()
ax.grid(True, alpha=0.3)
# Precision
ax = axes[1, 0]
if 'train_precision' in history and history['train_precision']:
ax.plot(epochs, history['train_precision'], 'b-', label='Train', linewidth=2)
if 'val_precision' in history and history['val_precision']:
ax.plot(epochs, history['val_precision'], 'r-', label='Val', linewidth=2)
ax.set_xlabel('Epoch')
ax.set_ylabel('Precision')
ax.set_title('Precision')
ax.legend()
ax.grid(True, alpha=0.3)
# Recall
ax = axes[1, 1]
if 'train_recall' in history and history['train_recall']:
ax.plot(epochs, history['train_recall'], 'b-', label='Train', linewidth=2)
if 'val_recall' in history and history['val_recall']:
ax.plot(epochs, history['val_recall'], 'r-', label='Val', linewidth=2)
ax.set_xlabel('Epoch')
ax.set_ylabel('Recall')
ax.set_title('Recall')
ax.legend()
ax.grid(True, alpha=0.3)
# Summary metrics bar chart
ax = axes[1, 2]
if history.get('val_iou') and history.get('val_dice'):
metrics = ['IoU', 'Dice', 'Precision', 'Recall']
final_values = [
history['val_iou'][-1] if history['val_iou'] else 0,
history['val_dice'][-1] if history['val_dice'] else 0,
history['val_precision'][-1] if history.get('val_precision') else 0,
history['val_recall'][-1] if history.get('val_recall') else 0
]
colors = sns.color_palette("husl", 4)
bars = ax.bar(metrics, final_values, color=colors)
ax.set_ylabel('Score')
ax.set_title('Final Validation Metrics')
ax.set_ylim(0, 1)
# Add value labels
for bar, val in zip(bars, final_values):
ax.text(bar.get_x() + bar.get_width()/2, bar.get_height() + 0.02,
f'{val:.3f}', ha='center', fontsize=10)
plt.tight_layout()
plt.savefig(save_path, dpi=150, bbox_inches='tight')
plt.close()
print(f"Training curves saved to {save_path}")
def plot_confusion_matrix(cm: np.ndarray,
class_names: List[str],
save_path: str,
title: str = "Confusion Matrix"):
"""
Plot confusion matrix
Args:
cm: Confusion matrix
class_names: Class names
save_path: Path to save plot
title: Plot title
"""
set_style()
fig, ax = plt.subplots(figsize=(8, 6))
sns.heatmap(cm, annot=True, fmt='d', cmap='Blues',
xticklabels=class_names,
yticklabels=class_names,
ax=ax)
ax.set_xlabel('Predicted')
ax.set_ylabel('True')
ax.set_title(title)
plt.tight_layout()
plt.savefig(save_path, dpi=150, bbox_inches='tight')
plt.close()
print(f"Confusion matrix saved to {save_path}")
def plot_feature_importance(importance: List[tuple],
save_path: str,
title: str = "Feature Importance"):
"""
Plot feature importance
Args:
importance: List of (feature_name, importance) tuples
save_path: Path to save plot
title: Plot title
"""
set_style()
fig, ax = plt.subplots(figsize=(10, 8))
names = [item[0] for item in importance]
values = [item[1] for item in importance]
colors = sns.color_palette("viridis", len(importance))
y_pos = np.arange(len(names))
ax.barh(y_pos, values, color=colors)
ax.set_yticks(y_pos)
ax.set_yticklabels(names)
ax.invert_yaxis()
ax.set_xlabel('Importance (Gain)')
ax.set_title(title)
plt.tight_layout()
plt.savefig(save_path, dpi=150, bbox_inches='tight')
plt.close()
print(f"Feature importance saved to {save_path}")
def plot_dataset_comparison(all_histories: Dict[str, Dict],
save_path: str):
"""
Plot comparison across datasets
Args:
all_histories: Dictionary of {dataset_name: history}
save_path: Path to save plot
"""
set_style()
fig, axes = plt.subplots(1, 2, figsize=(12, 5))
metrics = ['val_dice', 'val_iou']
titles = ['Validation Dice Score', 'Validation IoU']
for ax, metric, title in zip(axes, metrics, titles):
for dataset_name, history in all_histories.items():
if metric in history and history[metric]:
epochs = range(1, len(history[metric]) + 1)
ax.plot(epochs, history[metric], label=dataset_name, linewidth=2)
ax.set_xlabel('Epoch')
ax.set_ylabel(metric.replace('val_', '').replace('_', ' ').title())
ax.set_title(title)
ax.legend()
ax.grid(True, alpha=0.3)
plt.tight_layout()
plt.savefig(save_path, dpi=150, bbox_inches='tight')
plt.close()
print(f"Dataset comparison saved to {save_path}")
def plot_chunked_training_progress(chunk_histories: List[Dict],
save_path: str,
title: str = "Chunked Training Progress"):
"""
Plot progress across training chunks
Args:
chunk_histories: List of history dictionaries per chunk
save_path: Path to save plot
title: Plot title
"""
set_style()
fig, axes = plt.subplots(2, 2, figsize=(12, 10))
fig.suptitle(title, fontsize=14, fontweight='bold')
colors = sns.color_palette("husl", len(chunk_histories))
metrics = [
('train_loss', 'val_loss', 'Loss'),
('train_dice', 'val_dice', 'Dice Score'),
('train_iou', 'val_iou', 'IoU'),
('train_precision', 'val_precision', 'Precision')
]
for ax, (train_key, val_key, ylabel) in zip(axes.flat, metrics):
total_epochs = 0
for i, history in enumerate(chunk_histories):
if train_key in history and history[train_key]:
epochs = range(total_epochs + 1, total_epochs + len(history[train_key]) + 1)
ax.plot(epochs, history[train_key], '--', color=colors[i], alpha=0.5)
total_epochs += len(history[train_key])
total_epochs = 0
for i, history in enumerate(chunk_histories):
if val_key in history and history[val_key]:
epochs = range(total_epochs + 1, total_epochs + len(history[val_key]) + 1)
ax.plot(epochs, history[val_key], '-', color=colors[i],
label=f'Chunk {i+1}', linewidth=2)
# Add vertical line for chunk boundary
if i < len(chunk_histories) - 1:
ax.axvline(x=total_epochs + len(history[val_key]),
color='gray', linestyle=':', alpha=0.5)
total_epochs += len(history[val_key])
ax.set_xlabel('Epoch')
ax.set_ylabel(ylabel)
ax.set_title(f'Validation {ylabel}')
ax.legend()
ax.grid(True, alpha=0.3)
plt.tight_layout()
plt.savefig(save_path, dpi=150, bbox_inches='tight')
plt.close()
print(f"Chunked training progress saved to {save_path}")
def generate_training_report(history: Dict,
save_path: str,
dataset_name: str = "unknown"):
"""
Generate training report as text file
Args:
history: Training history
save_path: Path to save report
dataset_name: Dataset name
"""
with open(save_path, 'w') as f:
f.write("="*60 + "\n")
f.write(f"Training Report - {dataset_name}\n")
f.write("="*60 + "\n\n")
num_epochs = len(history.get('train_loss', []))
f.write(f"Total Epochs: {num_epochs}\n\n")
f.write("Final Metrics:\n")
f.write("-"*40 + "\n")
for key, values in history.items():
if values and isinstance(values, list):
final_value = values[-1]
if isinstance(final_value, (int, float)):
f.write(f" {key}: {final_value:.4f}\n")
f.write("\n")
f.write("Best Metrics:\n")
f.write("-"*40 + "\n")
for key, values in history.items():
if values and isinstance(values, list):
if 'loss' in key:
best_value = min(values)
best_epoch = values.index(best_value) + 1
else:
best_value = max(values)
best_epoch = values.index(best_value) + 1
if isinstance(best_value, (int, float)):
f.write(f" {key}: {best_value:.4f} (epoch {best_epoch})\n")
print(f"Training report saved to {save_path}")
|