File size: 4,626 Bytes
1ea7ba6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import json
import random
import time
from pathlib import Path

import numpy as np
import torch
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.metrics import (
    classification_report, confusion_matrix,
    accuracy_score, f1_score,
)

import config


def set_seed(seed: int = config.RANDOM_SEED):
    random.seed(seed)
    np.random.seed(seed)
    torch.manual_seed(seed)
    torch.backends.cudnn.deterministic = True
    torch.backends.cudnn.benchmark = False


def topk_accuracy(output: torch.Tensor, target: torch.Tensor, k: int = 5) -> float:
    with torch.no_grad():
        batch = target.size(0)
        _, pred = output.topk(k, dim=1, largest=True, sorted=True)
        correct = pred.eq(target.view(-1, 1).expand_as(pred))
        return correct.any(dim=1).float().sum().item() / batch


@torch.no_grad()
def measure_inference_time(model, loader, device, n_batches: int = 20) -> float:
    """Returns mean inference time in ms per image. Works with both SpiceFusionNet

    (uses forward_image) and plain timm models (uses __call__)."""
    model.eval()
    forward_fn = getattr(model, "forward_image", model)
    times = []
    for i, batch in enumerate(loader):
        if i >= n_batches:
            break
        imgs = batch[0].to(device)
        if device.type == "cuda":
            torch.cuda.synchronize()
        t0 = time.perf_counter()
        _ = forward_fn(imgs)
        if device.type == "cuda":
            torch.cuda.synchronize()
        elapsed = (time.perf_counter() - t0) * 1000  # ms
        times.append(elapsed / imgs.size(0))
    return float(np.mean(times))


def plot_training_curves(history: dict, output_dir: Path, prefix: str = ""):
    fig, axes = plt.subplots(1, 3, figsize=(18, 5))
    epochs = range(1, len(history["train_loss"]) + 1)

    axes[0].plot(epochs, history["train_loss"], label="train")
    axes[0].plot(epochs, history["val_loss"],   label="val")
    axes[0].set_title("Loss"); axes[0].set_xlabel("Epoch"); axes[0].legend()

    axes[1].plot(epochs, history["train_acc"], label="train")
    axes[1].plot(epochs, history["val_acc"],   label="val")
    axes[1].set_title("Accuracy"); axes[1].set_xlabel("Epoch"); axes[1].legend()

    axes[2].plot(epochs, history["lr"])
    axes[2].set_title("Learning Rate"); axes[2].set_xlabel("Epoch"); axes[2].set_yscale("log")

    plt.tight_layout()
    name = f"{prefix}_training_curves.png" if prefix else "training_curves.png"
    plt.savefig(output_dir / name, dpi=150)
    plt.close()


def plot_confusion_matrix(y_true, y_pred, classes, output_dir: Path, prefix: str = ""):
    cm      = confusion_matrix(y_true, y_pred)
    cm_norm = cm.astype(float) / cm.sum(axis=1, keepdims=True)

    fig, axes = plt.subplots(1, 2, figsize=(22, 9))
    for ax, data, title, fmt in [
        (axes[0], cm,      "Counts",     "d"),
        (axes[1], cm_norm, "Normalized", ".2f"),
    ]:
        sns.heatmap(data, annot=True, fmt=fmt, cmap="Blues",
                    xticklabels=classes, yticklabels=classes, ax=ax, linewidths=0.5)
        ax.set_title(title, fontsize=13)
        ax.set_xlabel("Predicted"); ax.set_ylabel("True")
        ax.tick_params(axis="x", rotation=45)

    plt.tight_layout()
    name = f"{prefix}_confusion_matrix.png" if prefix else "confusion_matrix.png"
    plt.savefig(output_dir / name, dpi=150)
    plt.close()


def compute_and_print_metrics(y_true, y_pred, classes, top5_acc=None, infer_ms=None) -> dict:
    acc  = accuracy_score(y_true, y_pred)
    f1_w = f1_score(y_true, y_pred, average="weighted")
    f1_m = f1_score(y_true, y_pred, average="macro")
    report = classification_report(y_true, y_pred, target_names=classes, digits=4)

    print(f"\n{'='*50}")
    print(f"  Top-1 Accuracy   : {acc:.4f}")
    if top5_acc is not None:
        print(f"  Top-5 Accuracy   : {top5_acc:.4f}")
    print(f"  Weighted F1      : {f1_w:.4f}")
    print(f"  Macro F1         : {f1_m:.4f}")
    if infer_ms is not None:
        print(f"  Inference time   : {infer_ms:.2f} ms/image")
    print(f"{'='*50}")
    print(f"\n{report}")

    metrics = {"top1_accuracy": acc, "f1_weighted": f1_w, "f1_macro": f1_m}
    if top5_acc  is not None: metrics["top5_accuracy"]    = top5_acc
    if infer_ms  is not None: metrics["inference_ms"]     = infer_ms
    return metrics


def save_metrics(metrics: dict, path: Path):
    with open(path, "w") as f:
        json.dump(metrics, f, indent=2)
    print(f"Metrics saved -> {path}")