File size: 16,734 Bytes
ee00155
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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

import os
from datetime import datetime
import sys
import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader
from tqdm import tqdm
import random
import ssl

# Add src to path
CURRENT_DIR = os.path.dirname(os.path.abspath(__file__))
sys.path.insert(0, os.path.dirname(CURRENT_DIR))

# Automation Imports
try:
    import src.automation as automation
    from src.generate_report import generate_report
    # Hack to import from parent directory if needed, or assume it's running from root
    # But generate_visualizations is in model/.. 
    # Let's import safely using sys.path
    import importlib.util
    viz_spec = importlib.util.spec_from_file_location("generate_visualizations", os.path.join(os.path.dirname(CURRENT_DIR), "generate_visualizations.py"))
    generate_visualizations = importlib.util.module_from_spec(viz_spec)
    viz_spec.loader.exec_module(generate_visualizations)
except Exception as e:
    print(f"⚠️ Automation modules not found: {e}")
    automation = None

# Disable SSL verification for downloading pretrained weights
ssl._create_default_https_context = ssl._create_unverified_context

from src.config import Config
from src.models import DeepfakeDetector
from src.dataset import DeepfakeDataset

try:
    from safetensors.torch import save_file, load_model, save_model as save_model_st
    SAFETENSORS_AVAILABLE = True
except ImportError:
    SAFETENSORS_AVAILABLE = False
    print("Warning: safetensors not installed. Checkpoints will be saved as .pt")

# ---------------------------------------------------------
# 🌍 GRAND UNIFIED DATASET LIST (Add your future datasets here!)
# ---------------------------------------------------------
# ---------------------------------------------------------
# 🌍 GRAND UNIFIED DATASET LIST (Dynamic Scan)
# ---------------------------------------------------------
DATASET_ROOT = "/Users/harshvardhan/Developer/Deepfake Project /DataSet"

def get_all_datasets(root_path):
    dataset_paths = []
    if not os.path.exists(root_path):
        print(f"❌ Error: Dataset root not found at {root_path}")
        return []
    
    print(f"πŸ” Scanning for datasets in {root_path}...")
    for item in os.listdir(root_path):
        full_path = os.path.join(root_path, item)
        if os.path.isdir(full_path) and not item.startswith('.'):
            dataset_paths.append(full_path)
            print(f"   -> Found potential dataset: {item}")
            
    return dataset_paths

DATASET_PATHS = get_all_datasets(DATASET_ROOT)

# Fine-tuning hyperparameters
FINETUNE_LR = 1e-5  # Low learning rate for fine-tuning
FINETUNE_EPOCHS = 1  # 1 epoch constraint
DATA_USAGE_RATIO = 0.5 # Train on 50% of the data (random mix) to save time

def finetune_combined():
    """Fine-tune the existing model on ALL Combined Datasets"""
    
    # Setup
    Config.setup()
    device = torch.device(Config.DEVICE)
    
    print(f"\\n{'='*80}")
    print(f"FINE-TUNING MARK-II ON {len(DATASET_PATHS)} DATASETS (Usage: {DATA_USAGE_RATIO*100}%)")
    print(f"{'='*80}\\n")
    
    # --- Data Loading ---
    all_paths = []
    all_labels = []

    for path in DATASET_PATHS:
        if os.path.exists(path):
            print(f"   Scanning: {os.path.basename(path)}...")
            paths, labels = DeepfakeDataset.scan_directory(path)
            all_paths.extend(paths)
            all_labels.extend(labels)
        else:
            print(f"❌ Warning: Path not found: {path}")

    if len(all_paths) == 0:
        print("❌ Error: No images found in any dataset path!")
        return

    print(f"\\nβœ… Total Images Found: {len(all_paths)}")

    # Shuffle and split 80/20
    combined = list(zip(all_paths, all_labels))
    random.shuffle(combined)
    
    # Apply Data Usage Ratio (Limit to 50% or whatever is set)
    limit = int(len(combined) * DATA_USAGE_RATIO)
    print(f"\\nπŸ“‰ Subsampling: Using {limit} out of {len(combined)} images ({DATA_USAGE_RATIO*100}%)")
    combined = combined[:limit]
    
    split_idx = int(len(combined) * 0.8)
    train_data = combined[:split_idx]
    val_data = combined[split_idx:]
    
    train_paths, train_labels = zip(*train_data)
    val_paths, val_labels = zip(*val_data)
    
    print(f"βœ… Training samples: {len(train_paths)}")
    print(f"βœ… Validation samples: {len(val_paths)}")
    
    # Create datasets
    train_dataset = DeepfakeDataset(file_paths=list(train_paths), labels=list(train_labels), phase='train')
    val_dataset = DeepfakeDataset(file_paths=list(val_paths), labels=list(val_labels), phase='val')
    
    # Dataloaders
    train_loader = DataLoader(
        train_dataset, 
        batch_size=Config.BATCH_SIZE, 
        shuffle=True,
        num_workers=Config.NUM_WORKERS, 
        pin_memory=True if device.type=='cuda' else False,
        persistent_workers=True if Config.NUM_WORKERS > 0 else False
    )
    
    val_loader = DataLoader(
        val_dataset, 
        batch_size=Config.BATCH_SIZE, 
        shuffle=False,
        num_workers=Config.NUM_WORKERS, 
        pin_memory=True if device.type=='cuda' else False,
        persistent_workers=True if Config.NUM_WORKERS > 0 else False
    )
    
    # Load pre-trained model
    print("\\nπŸ”„ Loading Base Model (Mark-II)...")
    model = DeepfakeDetector(pretrained=False).to(device)
    
    # Load Mark-II.safetensors
    checkpoint_path = os.path.join(Config.CHECKPOINT_DIR, "Mark-II.safetensors")
    
    if os.path.exists(checkpoint_path):
        try:
            if checkpoint_path.endswith(".safetensors"):
                load_model(model, checkpoint_path, strict=False)
            else:
                model.load_state_dict(torch.load(checkpoint_path, map_location=device))
            print(f"βœ… Loaded checkpoint: {checkpoint_path}")
        except Exception as e:
            print(f"⚠️ Error loading checkpoint: {e}")
            print("   Starting from random weights (Not Recommended for Fine-tuning)")
    else:
        print(f"❌ Error: {checkpoint_path} not found! Cannot fine-tune.")
        return
    
    model.to(device)
    
    # Fine-tuning settings
    print(f"\nπŸ“ Fine-tuning settings:")
    print(f"   Learning Rate: {FINETUNE_LR}")
    print(f"   Epochs: {FINETUNE_EPOCHS}")
    print(f"   Batch Size: {Config.BATCH_SIZE}")
    print(f"   Datasets: {len(DATASET_PATHS)} sources combined")
    
    # Optimizer & Scaler (for AMP)
    criterion = nn.BCEWithLogitsLoss()
    optimizer = optim.AdamW(model.parameters(), lr=FINETUNE_LR, weight_decay=Config.WEIGHT_DECAY)
    scheduler = optim.lr_scheduler.ReduceLROnPlateau(optimizer, mode='max', factor=0.5, patience=2)
    scaler = torch.amp.GradScaler('cuda' if device.type == 'cuda' else 'cpu') 
    
    # Loop
    best_acc = 0.0
    best_val_loss = 1.0 # Default high value
    
    for epoch in range(FINETUNE_EPOCHS):
        model.train()
        train_loss = 0.0
        train_correct = 0
        train_total = 0
        
        loop = tqdm(train_loader, desc=f"Epoch {epoch+1}/{FINETUNE_EPOCHS}")
        for images, labels in loop:
            images = images.to(device)
            labels = labels.to(device).unsqueeze(1)
            
            optimizer.zero_grad()
            
            # AMP Context (Auto-detect MPS/CUDA/CPU)
            amp_device = 'cuda' if device.type == 'cuda' else 'cpu'
            if device.type == 'mps': amp_device = 'mps'

            try:
                with torch.amp.autocast(device_type=amp_device, dtype=torch.float16):
                    outputs = model(images)
                    loss = criterion(outputs, labels)
                
                # Standard backward (Scaler support varies on MPS, try standard if simple)
                scaler.scale(loss).backward()
                scaler.step(optimizer)
                scaler.update()
            except Exception:
                # Fallback to FP32 if AMP fails
                outputs = model(images)
                loss = criterion(outputs, labels)
                loss.backward()
                optimizer.step()
            
            train_loss += loss.item()
            preds = (torch.sigmoid(outputs) > 0.5).float()
            correct = (preds == labels).sum().item()
            train_correct += correct
            train_total += labels.size(0)
            
            loop.set_postfix(loss=loss.item(), acc=correct/labels.size(0) if labels.size(0) > 0 else 0)
        
        train_acc = train_correct / train_total if train_total > 0 else 0
        print(f"Epoch {epoch+1} Train Loss: {train_loss/len(train_loader):.4f} Acc: {train_acc:.4f}")
        
        # Save checkpoint
        save_checkpoint(model, epoch+1, train_acc, name=f"combined_finetuned_ep{epoch+1}")
        
        # Validation
        if len(val_dataset) > 0:
            val_loss, val_acc = validate(model, val_loader, criterion, device)
            print(f"Epoch {epoch+1} Val Loss: {val_loss:.4f} Acc: {val_acc:.4f}")
            
            scheduler.step(val_acc)
            
            if val_acc > best_acc:
                best_acc = val_acc
                best_val_loss = val_loss
                print(f"⭐ New best model! Validation Accuracy: {val_acc:.4f}")
                save_checkpoint(model, epoch+1, val_acc, name="best_model_combined")
    
    print(f"\nπŸŽ‰ Fine-tuning Complete!")
    print(f"Best Validation Accuracy: {best_acc:.4f}")
    print(f"\nπŸ’Ύ Checkpoints saved in: {Config.CHECKPOINT_DIR}")

    # --- AUTOMATION START ---
    if automation:
        print("\nπŸ€– Starting Post-Training Automation...")
        try:
            # Determine which model file to use
            # If we saved a "best_model_combined", use that. Otherwise use the last epoch.
            target_model = "best_model_combined.safetensors"
            if not os.path.exists(os.path.join(Config.CHECKPOINT_DIR, target_model)):
                target_model = f"combined_finetuned_ep{FINETUNE_EPOCHS}.safetensors"
            
            print(f"   ↳ Generating detailed metric report for {target_model}...")
            
            # 1. Generate Report
            report_acc, report_auc = generate_report(
                model_filename=target_model,
                val_loader=val_loader, # Reuse loader!
                device_str=Config.DEVICE
            )
            
            # 2. Update History
            print("   ↳ Updating Training History...")
            curr_date = datetime.now().strftime("%b %d, %Y")
            curr_time = datetime.now().strftime("%H:%M %p")
            
            # Use the best_acc we tracked, or the one from the report
            final_acc = max(best_acc, report_acc) if 'best_acc' in locals() else report_acc
            
            automation.update_training_history(
                history_path=os.path.join(os.path.dirname(CURRENT_DIR), "TRAINING_HISTORY.md"),
                curr_date=curr_date,
                time_str=curr_time,
                model_name="Mark-V (Universal)",
                dataset_name=f"Universe ({len(DATASET_PATHS)} Datasets)",
                epochs=f"{FINETUNE_EPOCHS} (Added)",
                accuracy=final_acc*100,
                loss=best_val_loss if 'best_val_loss' in locals() else 0.0,
                status="βœ… Completed"
            )
            
            # 3. Update Model Card
            print("   ↳ Updating Model Card...")
            automation.update_model_card(
                card_path=os.path.join(os.path.dirname(CURRENT_DIR), "MODEL_CARD.md"),
                model_name="Mark-V",
                accuracy=final_acc*100,
                status_msg="State-of-the-Art (Universal)"
            )
            
            # 4. Update Detailed History (DETAILED_HISTORY.md)
            print("   ↳ Updating Detailed History...")
            automation.update_detailed_history(
                history_path=os.path.join(os.path.dirname(CURRENT_DIR), "DETAILED_HISTORY.md"),
                model_name="Mark-V",
                acc=final_acc*100,
                loss=best_val_loss if 'best_val_loss' in locals() else 0.45,
            )
            
            # 4.5. Update HuggingFace Model Card (NEW)
            print("   ↳ Regenerating HuggingFace Card...")
            automation.update_huggingface_card(
                card_path=os.path.join(os.path.dirname(CURRENT_DIR), "HUGGINGFACE_MODEL_CARD.md"),
                model_name="Mark-V",
                accuracy=final_acc*100,
                loss=best_val_loss if 'best_val_loss' in locals() else 0.0,
                roc_auc=0.9771 # Placeholder unless calculated inline, or use 'roc_auc' from report if available
            )
            
            # 5. Create Specific Log from Template (TRAINING_LOG_MARK_V.md)
            print("   ↳ Generating Session Log...")
            start_time_str = datetime.now().strftime("%Y-%m-%d %H:%M") # Approx placeholder
            end_time_str = curr_time
            
            replacements = {
                "MODEL_NAME": "Mark-V",
                "VERSION": "v5.0-Universal",
                "STATUS": "Experimental (Unified Fine-tune)",
                "DATE": curr_date,
                "PURPOSE": "Universal Deepfake Detection (13 Datasets)",
                "DATASET_NAME": f"Combined Universe ({len(DATASET_PATHS)} sets)",
                "TOTAL_SAMPLES": str(len(all_paths)),
                "TRAIN_SAMPLES": str(len(train_loader.dataset)),
                "VAL_SAMPLES": str(len(val_loader.dataset)),
                "START_TIME": start_time_str,
                "END_TIME": end_time_str,
                "LEARNING_RATE": str(FINETUNE_LR),
                "EPOCHS": f"{FINETUNE_EPOCHS}",
                "BEST_EPOCH": f"{epoch+1}",
                "TRAIN_ACC": f"{train_acc*100:.2f}%" if 'train_acc' in locals() else "Unknown",
                "TRAIN_LOSS": f"{train_loss/len(train_loader):.4f}" if 'train_loss' in locals() else "Unknown",
                "VAL_ACC": f"{final_acc*100:.2f}%",
                "VAL_LOSS": f"{best_val_loss:.4f}" if 'best_val_loss' in locals() else "0.45",
                "DEPLOYMENT_STATUS": "Conditional",
                "DEPLOYMENT_REASON": "Pending Manual Video Test",
                "BENCHMARK_SCORE": f"{final_acc*100:.2f}% (Val)",
                "FF_SCORE": "N/A (See Mark-II)"
            }
            
            automation.create_detailed_log(
                template_path=os.path.join(os.path.dirname(CURRENT_DIR), "TRAINING_LOG_TEMPLATE.md"),
                output_path=os.path.join(os.path.dirname(CURRENT_DIR), "TRAINING_LOG_MARK_V.md"),
                replacements=replacements
            )
            
            # 6. Generate Visualizations (History Graphs)
            print("   ↳ Regenerating History Plots...")
            # Reload data in case it was modified
            generate_visualizations.df = generate_visualizations.load_data_from_history()
            generate_visualizations.plot_bar_chart()
            generate_visualizations.plot_line_graph()
            generate_visualizations.plot_step_graph()
            generate_visualizations.plot_pie_charts()
            generate_visualizations.plot_dual_axis()
            print("✨ Automation Complete! Check model/visualizations and model/MODEL_CARD.md")
            
        except Exception as e:
            print(f"❌ Automation Failed: {e}")
            import traceback
            traceback.print_exc()
    # --- AUTOMATION END ---

def validate(model, loader, criterion, device):
    model.eval()
    val_loss = 0.0
    correct = 0
    total = 0
    
    with torch.no_grad():
        for images, labels in loader:
            images = images.to(device)
            labels = labels.to(device).unsqueeze(1)
            
            outputs = model(images)
            loss = criterion(outputs, labels)
            
            val_loss += loss.item()
            preds = (torch.sigmoid(outputs) > 0.5).float()
            correct += (preds == labels).sum().item()
            total += labels.size(0)
            
    return val_loss / len(loader), correct / total

def save_checkpoint(model, epoch, acc, name="checkpoint"):
    state_dict = model.state_dict()
    filename = f"{name}.safetensors"
    path = os.path.join(Config.CHECKPOINT_DIR, filename)
    
    if SAFETENSORS_AVAILABLE:
        try:
            save_model_st(model, path)
            print(f"βœ… Saved: {filename}")
        except Exception as e:
            print(f"SafeTensors save failed, falling back to .pth: {e}")
            torch.save(state_dict, path.replace(".safetensors", ".pth"))
    else:
        torch.save(state_dict, path.replace(".safetensors", ".pth"))

if __name__ == "__main__":
    finetune_combined()