File size: 16,618 Bytes
3f891b2 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 |
"""
PlainMLP vs ResMLP Comparison - FAIR INITIALIZATION VERSION
This version uses IDENTICAL initialization for both models to ensure
a fair comparison. Both use:
- Kaiming He initialization
- Weight scaling by 1/sqrt(num_layers)
- Zero bias initialization
The ONLY difference is the residual connection: x = x + f(x) vs x = f(x)
"""
import torch
import torch.nn as nn
import numpy as np
import matplotlib.pyplot as plt
from typing import Dict, List, Tuple
import json
import os
# Set random seeds for reproducibility
torch.manual_seed(42)
np.random.seed(42)
# Configuration
NUM_LAYERS = 20
HIDDEN_DIM = 64
NUM_SAMPLES = 1024
TRAINING_STEPS = 500
LEARNING_RATE = 1e-3
BATCH_SIZE = 64
print(f"[Config] Layers: {NUM_LAYERS}, Hidden Dim: {HIDDEN_DIM}")
print(f"[Config] Samples: {NUM_SAMPLES}, Steps: {TRAINING_STEPS}, LR: {LEARNING_RATE}")
print(f"[Config] FAIR COMPARISON: Both models use identical initialization")
class PlainMLP(nn.Module):
"""Plain MLP: x = ReLU(Linear(x)) for each layer
NOW WITH SAME INITIALIZATION AS ResMLP:
- Kaiming He initialization
- Weight scaling by 1/sqrt(num_layers)
- Zero bias
"""
def __init__(self, dim: int, num_layers: int):
super().__init__()
self.layers = nn.ModuleList()
self.num_layers = num_layers
for _ in range(num_layers):
layer = nn.Linear(dim, dim)
# SAME initialization as ResMLP
nn.init.kaiming_normal_(layer.weight, mode='fan_in', nonlinearity='relu')
layer.weight.data *= 1.0 / np.sqrt(num_layers) # Same scaling!
nn.init.zeros_(layer.bias)
self.layers.append(layer)
self.activation = nn.ReLU()
def forward(self, x: torch.Tensor) -> torch.Tensor:
for layer in self.layers:
x = self.activation(layer(x)) # NO residual connection
return x
class ResMLP(nn.Module):
"""Residual MLP: x = x + ReLU(Linear(x)) for each layer
Uses same initialization as PlainMLP:
- Kaiming He initialization
- Weight scaling by 1/sqrt(num_layers)
- Zero bias
"""
def __init__(self, dim: int, num_layers: int):
super().__init__()
self.layers = nn.ModuleList()
self.num_layers = num_layers
for _ in range(num_layers):
layer = nn.Linear(dim, dim)
# Same initialization as PlainMLP
nn.init.kaiming_normal_(layer.weight, mode='fan_in', nonlinearity='relu')
layer.weight.data *= 1.0 / np.sqrt(num_layers) # Same scaling
nn.init.zeros_(layer.bias)
self.layers.append(layer)
self.activation = nn.ReLU()
def forward(self, x: torch.Tensor) -> torch.Tensor:
for layer in self.layers:
x = x + self.activation(layer(x)) # WITH residual connection
return x
def generate_identity_data(num_samples: int, dim: int) -> Tuple[torch.Tensor, torch.Tensor]:
"""Generate synthetic data where Y = X, with X ~ U(-1, 1)"""
X = torch.empty(num_samples, dim).uniform_(-1, 1)
Y = X.clone()
return X, Y
def train_model(model: nn.Module, X: torch.Tensor, Y: torch.Tensor,
steps: int, lr: float, batch_size: int) -> List[float]:
"""Train model and record loss at each step"""
optimizer = torch.optim.Adam(model.parameters(), lr=lr)
criterion = nn.MSELoss()
losses = []
num_samples = X.shape[0]
for step in range(steps):
# Random batch sampling
indices = torch.randint(0, num_samples, (batch_size,))
batch_x = X[indices]
batch_y = Y[indices]
# Forward pass
optimizer.zero_grad()
output = model(batch_x)
loss = criterion(output, batch_y)
# Backward pass
loss.backward()
optimizer.step()
losses.append(loss.item())
if step % 100 == 0:
print(f" Step {step}/{steps}, Loss: {loss.item():.6f}")
return losses
class ActivationGradientHook:
"""Hook to capture activations and gradients at each layer"""
def __init__(self):
self.activations: List[torch.Tensor] = []
self.gradients: List[torch.Tensor] = []
self.handles = []
def register_hooks(self, model: nn.Module):
"""Register forward and backward hooks on each layer"""
for layer in model.layers:
handle_fwd = layer.register_forward_hook(self._forward_hook)
handle_bwd = layer.register_full_backward_hook(self._backward_hook)
self.handles.extend([handle_fwd, handle_bwd])
def _forward_hook(self, module, input, output):
self.activations.append(output.detach().clone())
def _backward_hook(self, module, grad_input, grad_output):
self.gradients.append(grad_output[0].detach().clone())
def clear(self):
self.activations = []
self.gradients = []
def remove_hooks(self):
for handle in self.handles:
handle.remove()
self.handles = []
def get_activation_stats(self) -> Tuple[List[float], List[float]]:
"""Get mean and std of activations for each layer"""
means = [act.mean().item() for act in self.activations]
stds = [act.std().item() for act in self.activations]
return means, stds
def get_gradient_norms(self) -> List[float]:
"""Get L2 norm of gradients for each layer (in forward order)"""
norms = [grad.norm(2).item() for grad in reversed(self.gradients)]
return norms
def analyze_final_state(model: nn.Module, dim: int, batch_size: int = 64) -> Dict:
"""Perform forward/backward pass and capture activation/gradient stats"""
hook = ActivationGradientHook()
hook.register_hooks(model)
# Generate new random batch
X_test = torch.empty(batch_size, dim).uniform_(-1, 1)
Y_test = X_test.clone()
# Forward pass
model.zero_grad()
output = model(X_test)
loss = nn.MSELoss()(output, Y_test)
# Backward pass
loss.backward()
# Get statistics
act_means, act_stds = hook.get_activation_stats()
grad_norms = hook.get_gradient_norms()
hook.remove_hooks()
return {
'activation_means': act_means,
'activation_stds': act_stds,
'gradient_norms': grad_norms,
'final_loss': loss.item()
}
def plot_training_loss(plain_losses: List[float], res_losses: List[float], save_path: str):
"""Plot training loss curves for both models"""
fig, ax = plt.subplots(figsize=(10, 6))
steps = range(len(plain_losses))
ax.plot(steps, plain_losses, label='PlainMLP (20 layers)', color='#e74c3c',
alpha=0.8, linewidth=2)
ax.plot(steps, res_losses, label='ResMLP (20 layers)', color='#3498db',
alpha=0.8, linewidth=2)
ax.set_xlabel('Training Steps', fontsize=12)
ax.set_ylabel('MSE Loss', fontsize=12)
ax.set_title('Training Loss: PlainMLP vs ResMLP (FAIR Initialization)\nIdentity Task (Y = X)', fontsize=14)
ax.legend(fontsize=11, loc='upper right')
ax.grid(True, alpha=0.3)
ax.set_yscale('log')
# Add final loss annotations
final_plain = plain_losses[-1]
final_res = res_losses[-1]
# Text box with final results
textstr = f'Final Loss:\n PlainMLP: {final_plain:.4f}\n ResMLP: {final_res:.4f}\n Improvement: {final_plain/final_res:.1f}x'
props = dict(boxstyle='round', facecolor='wheat', alpha=0.8)
ax.text(0.02, 0.02, textstr, transform=ax.transAxes, fontsize=10,
verticalalignment='bottom', bbox=props)
plt.tight_layout()
plt.savefig(save_path, dpi=150, bbox_inches='tight')
plt.close()
print(f"[Plot] Saved training loss plot to {save_path}")
def plot_gradient_magnitudes(plain_grads: List[float], res_grads: List[float], save_path: str):
"""Plot gradient magnitude vs layer depth"""
fig, ax = plt.subplots(figsize=(10, 6))
layers = range(1, len(plain_grads) + 1)
ax.plot(layers, plain_grads, 'o-', label='PlainMLP', color='#e74c3c',
markersize=8, linewidth=2, markeredgecolor='white', markeredgewidth=1)
ax.plot(layers, res_grads, 's-', label='ResMLP', color='#3498db',
markersize=8, linewidth=2, markeredgecolor='white', markeredgewidth=1)
ax.set_xlabel('Layer Depth (1 = first layer, 20 = last layer)', fontsize=12)
ax.set_ylabel('Gradient L2 Norm (log scale)', fontsize=12)
ax.set_title('Gradient Magnitude vs Layer Depth (Fair Initialization)', fontsize=14)
ax.legend(fontsize=11)
ax.grid(True, alpha=0.3)
ax.set_yscale('log')
plt.tight_layout()
plt.savefig(save_path, dpi=150, bbox_inches='tight')
plt.close()
print(f"[Plot] Saved gradient magnitude plot to {save_path}")
def plot_activation_means(plain_means: List[float], res_means: List[float], save_path: str):
"""Plot activation mean vs layer depth"""
fig, ax = plt.subplots(figsize=(10, 6))
layers = range(1, len(plain_means) + 1)
ax.plot(layers, plain_means, 'o-', label='PlainMLP', color='#e74c3c',
markersize=8, linewidth=2, markeredgecolor='white', markeredgewidth=1)
ax.plot(layers, res_means, 's-', label='ResMLP', color='#3498db',
markersize=8, linewidth=2, markeredgecolor='white', markeredgewidth=1)
ax.axhline(y=0, color='gray', linestyle='--', alpha=0.5, linewidth=1)
ax.set_xlabel('Layer Depth', fontsize=12)
ax.set_ylabel('Activation Mean', fontsize=12)
ax.set_title('Activation Mean vs Layer Depth (Fair Initialization)', fontsize=14)
ax.legend(fontsize=11)
ax.grid(True, alpha=0.3)
plt.tight_layout()
plt.savefig(save_path, dpi=150, bbox_inches='tight')
plt.close()
print(f"[Plot] Saved activation mean plot to {save_path}")
def plot_activation_stds(plain_stds: List[float], res_stds: List[float], save_path: str):
"""Plot activation std vs layer depth"""
fig, ax = plt.subplots(figsize=(10, 6))
layers = range(1, len(plain_stds) + 1)
ax.plot(layers, plain_stds, 'o-', label='PlainMLP', color='#e74c3c',
markersize=8, linewidth=2, markeredgecolor='white', markeredgewidth=1)
ax.plot(layers, res_stds, 's-', label='ResMLP', color='#3498db',
markersize=8, linewidth=2, markeredgecolor='white', markeredgewidth=1)
ax.set_xlabel('Layer Depth', fontsize=12)
ax.set_ylabel('Activation Standard Deviation', fontsize=12)
ax.set_title('Activation Std vs Layer Depth (Fair Initialization)', fontsize=14)
ax.legend(fontsize=11)
ax.grid(True, alpha=0.3)
plt.tight_layout()
plt.savefig(save_path, dpi=150, bbox_inches='tight')
plt.close()
print(f"[Plot] Saved activation std plot to {save_path}")
def main():
print("=" * 70)
print("PlainMLP vs ResMLP: FAIR COMPARISON (Identical Initialization)")
print("=" * 70)
# Ensure plots directory exists
os.makedirs('plots_fair', exist_ok=True)
# Generate synthetic data
print("\n[1] Generating synthetic identity data...")
X, Y = generate_identity_data(NUM_SAMPLES, HIDDEN_DIM)
print(f" Data shape: X={X.shape}, Y={Y.shape}")
print(f" X range: [{X.min():.3f}, {X.max():.3f}]")
print(f" Task: Learn Y = X (identity mapping)")
# Initialize models
print("\n[2] Initializing models with IDENTICAL initialization...")
print(" Both use: Kaiming He + 1/sqrt(num_layers) scaling + zero bias")
plain_mlp = PlainMLP(HIDDEN_DIM, NUM_LAYERS)
res_mlp = ResMLP(HIDDEN_DIM, NUM_LAYERS)
plain_params = sum(p.numel() for p in plain_mlp.parameters())
res_params = sum(p.numel() for p in res_mlp.parameters())
print(f" PlainMLP parameters: {plain_params:,}")
print(f" ResMLP parameters: {res_params:,}")
# Verify initialization is the same
print("\n Verifying initialization parity...")
plain_w_norm = sum(p.norm().item() for p in plain_mlp.parameters())
res_w_norm = sum(p.norm().item() for p in res_mlp.parameters())
print(f" PlainMLP total weight norm: {plain_w_norm:.4f}")
print(f" ResMLP total weight norm: {res_w_norm:.4f}")
# Train PlainMLP
print("\n[3] Training PlainMLP...")
plain_losses = train_model(plain_mlp, X, Y, TRAINING_STEPS, LEARNING_RATE, BATCH_SIZE)
print(f" Final loss: {plain_losses[-1]:.6f}")
# Train ResMLP
print("\n[4] Training ResMLP...")
res_losses = train_model(res_mlp, X, Y, TRAINING_STEPS, LEARNING_RATE, BATCH_SIZE)
print(f" Final loss: {res_losses[-1]:.6f}")
# Calculate improvement
improvement = plain_losses[-1] / res_losses[-1]
print(f"\n >>> ResMLP achieves {improvement:.1f}x lower loss than PlainMLP <<<")
# Final state analysis
print("\n[5] Analyzing final state of trained models...")
print(" Running forward/backward pass on new random batch...")
print(" Analyzing PlainMLP...")
plain_stats = analyze_final_state(plain_mlp, HIDDEN_DIM)
print(" Analyzing ResMLP...")
res_stats = analyze_final_state(res_mlp, HIDDEN_DIM)
# Print detailed analysis
print("\n[6] Detailed Analysis:")
print("\n === Loss Comparison ===")
print(f" PlainMLP - Initial: {plain_losses[0]:.4f}, Final: {plain_losses[-1]:.4f}")
print(f" ResMLP - Initial: {res_losses[0]:.4f}, Final: {res_losses[-1]:.4f}")
print("\n === Gradient Flow (L2 norms) ===")
print(f" PlainMLP - Layer 1: {plain_stats['gradient_norms'][0]:.2e}, Layer 20: {plain_stats['gradient_norms'][-1]:.2e}")
print(f" ResMLP - Layer 1: {res_stats['gradient_norms'][0]:.2e}, Layer 20: {res_stats['gradient_norms'][-1]:.2e}")
print("\n === Activation Statistics ===")
print(f" PlainMLP - Std range: [{min(plain_stats['activation_stds']):.4f}, {max(plain_stats['activation_stds']):.4f}]")
print(f" ResMLP - Std range: [{min(res_stats['activation_stds']):.4f}, {max(res_stats['activation_stds']):.4f}]")
# Generate plots
print("\n[7] Generating plots...")
plot_training_loss(plain_losses, res_losses, 'plots_fair/training_loss.png')
plot_gradient_magnitudes(plain_stats['gradient_norms'], res_stats['gradient_norms'],
'plots_fair/gradient_magnitude.png')
plot_activation_means(plain_stats['activation_means'], res_stats['activation_means'],
'plots_fair/activation_mean.png')
plot_activation_stds(plain_stats['activation_stds'], res_stats['activation_stds'],
'plots_fair/activation_std.png')
# Save results to JSON
results = {
'config': {
'num_layers': NUM_LAYERS,
'hidden_dim': HIDDEN_DIM,
'num_samples': NUM_SAMPLES,
'training_steps': TRAINING_STEPS,
'learning_rate': LEARNING_RATE,
'batch_size': BATCH_SIZE,
'initialization': 'Kaiming He + 1/sqrt(num_layers) scaling (IDENTICAL for both)'
},
'plain_mlp': {
'final_loss': plain_losses[-1],
'initial_loss': plain_losses[0],
'loss_history': plain_losses,
'gradient_norms': plain_stats['gradient_norms'],
'activation_means': plain_stats['activation_means'],
'activation_stds': plain_stats['activation_stds']
},
'res_mlp': {
'final_loss': res_losses[-1],
'initial_loss': res_losses[0],
'loss_history': res_losses,
'gradient_norms': res_stats['gradient_norms'],
'activation_means': res_stats['activation_means'],
'activation_stds': res_stats['activation_stds']
},
'summary': {
'loss_improvement': improvement,
'plain_grad_range': [min(plain_stats['gradient_norms']), max(plain_stats['gradient_norms'])],
'res_grad_range': [min(res_stats['gradient_norms']), max(res_stats['gradient_norms'])],
'plain_std_range': [min(plain_stats['activation_stds']), max(plain_stats['activation_stds'])],
'res_std_range': [min(res_stats['activation_stds']), max(res_stats['activation_stds'])]
}
}
with open('results_fair.json', 'w') as f:
json.dump(results, f, indent=2)
print("\n[8] Results saved to results_fair.json")
print("\n" + "=" * 70)
print("FAIR COMPARISON Experiment completed successfully!")
print("=" * 70)
return results
if __name__ == "__main__":
results = main()
|