File size: 4,197 Bytes
1f39ae1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
#!/usr/bin/env python3
"""
Test script to verify the training setup works correctly
"""

import torch
from torch.utils.data import DataLoader

from morphological_dataset import MorphologicalDataset, build_vocabulary, collate_fn
from transformer import TagTransformer, PAD_IDX, DEVICE

def test_training_setup():
    """Test the complete training setup"""
    
    print("=== Testing Training Setup ===")
    
    # Data file paths
    train_src = '10L_90NL/train/run1/train.10L_90NL_1_1.src'
    train_tgt = '10L_90NL/train/run1/train.10L_90NL_1_1.tgt'
    
    # Build vocabularies
    print("Building vocabularies...")
    src_vocab = build_vocabulary([train_src])
    tgt_vocab = build_vocabulary([train_tgt])
    
    print(f"Source vocabulary size: {len(src_vocab)}")
    print(f"Target vocabulary size: {len(tgt_vocab)}")
    
    # Count feature tokens
    feature_tokens = [token for token in src_vocab.keys() 
                     if token.startswith('<') and token.endswith('>')]
    nb_attr = len(feature_tokens)
    
    print(f"Number of feature tokens: {nb_attr}")
    print(f"Feature examples: {feature_tokens[:5]}")
    
    # Create dataset
    print("\nCreating dataset...")
    dataset = MorphologicalDataset(train_src, train_tgt, src_vocab, tgt_vocab, max_length=50)
    print(f"Dataset size: {len(dataset)}")
    
    # Test data loading
    print("\nTesting data loading...")
    sample_src, sample_tgt = dataset[0]
    print(f"Sample source: {' '.join(sample_src)}")
    print(f"Sample target: {' '.join(sample_tgt)}")
    
    # Create dataloader
    print("\nCreating dataloader...")
    dataloader = DataLoader(
        dataset, 
        batch_size=4, 
        shuffle=False, 
        collate_fn=lambda batch: collate_fn(batch, src_vocab, tgt_vocab, 50),
        num_workers=0
    )
    
    # Test batch loading
    print("\nTesting batch loading...")
    for batch_idx, (src, src_mask, tgt, tgt_mask) in enumerate(dataloader):
        print(f"Batch {batch_idx}:")
        print(f"  Source shape: {src.shape}")
        print(f"  Source mask shape: {src_mask.shape}")
        print(f"  Target shape: {tgt.shape}")
        print(f"  Target mask shape: {tgt_mask.shape}")
        
        if batch_idx >= 1:  # Only test first 2 batches
            break
    
    # Create model
    print("\nCreating model...")
    config = {
        'embed_dim': 64,  # Small for testing
        'nb_heads': 4,
        'src_hid_size': 128,
        'src_nb_layers': 2,
        'trg_hid_size': 128,
        'trg_nb_layers': 2,
        'dropout_p': 0.0,
        'tie_trg_embed': True,
        'label_smooth': 0.0,
    }
    
    model = TagTransformer(
        src_vocab_size=len(src_vocab),
        trg_vocab_size=len(tgt_vocab),
        embed_dim=config['embed_dim'],
        nb_heads=config['nb_heads'],
        src_hid_size=config['src_hid_size'],
        src_nb_layers=config['src_nb_layers'],
        trg_hid_size=config['trg_hid_size'],
        trg_nb_layers=config['trg_nb_layers'],
        dropout_p=config['dropout_p'],
        tie_trg_embed=config['tie_trg_embed'],
        label_smooth=config['label_smooth'],
        nb_attr=nb_attr,
        src_c2i=src_vocab,
        trg_c2i=tgt_vocab,
        attr_c2i={},
    )
    
    model = model.to(DEVICE)
    print(f"Model created with {model.count_nb_params():,} parameters")
    
    # Test forward pass
    print("\nTesting forward pass...")
    model.eval()
    
    with torch.no_grad():
        # Get a batch
        src, src_mask, tgt, tgt_mask = next(iter(dataloader))
        src, src_mask, tgt, tgt_mask = (
            src.to(DEVICE), src_mask.to(DEVICE), 
            tgt.to(DEVICE), tgt_mask.to(DEVICE)
        )
        
        # Forward pass
        output = model(src, src_mask, tgt, tgt_mask)
        print(f"  Output shape: {output.shape}")
        
        # Test loss computation
        loss = model.loss(output[:-1], tgt[1:])
        print(f"  Loss: {loss.item():.4f}")
    
    print("\n✓ Training setup test completed successfully!")
    print("\nReady to start training with:")
    print(f"python scripts/train_morphological.py --output_dir ./models")

if __name__ == '__main__':
    test_training_setup()