{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.11.11","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"nvidiaTeslaT4","dataSources":[{"sourceId":91844,"databundleVersionId":11361821,"sourceType":"competition"},{"sourceId":11053663,"sourceType":"datasetVersion","datasetId":6886569}],"dockerImageVersionId":31012,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import os\nimport logging\nimport random\nimport gc\nimport time\nimport cv2\nimport math\nimport warnings\nfrom pathlib import Path\n\nimport numpy as np\nimport pandas as pd\nfrom sklearn.model_selection import StratifiedKFold\nfrom sklearn.metrics import roc_auc_score\nimport librosa\n\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nimport torch.optim as optim\nfrom torch.optim import lr_scheduler\nfrom torch.utils.data import Dataset, DataLoader\n\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nfrom tqdm.auto import tqdm\n\nimport timm\n\nwarnings.filterwarnings(\"ignore\")\nlogging.basicConfig(level=logging.ERROR)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-19T11:12:39.079446Z","iopub.execute_input":"2025-04-19T11:12:39.079762Z","iopub.status.idle":"2025-04-19T11:12:53.193257Z","shell.execute_reply.started":"2025-04-19T11:12:39.079739Z","shell.execute_reply":"2025-04-19T11:12:53.192718Z"},"jupyter":{"source_hidden":true}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"#自定义类\nimport torch\nimport torch.nn as nn\n\nclass CBAM(nn.Module):\n    def __init__(self, channel, reduction=16, spatial_kernel=7):\n        super(CBAM, self).__init__()\n \n        # channel attention 压缩H,W为1\n        self.max_pool = nn.AdaptiveMaxPool2d(1)\n        self.avg_pool = nn.AdaptiveAvgPool2d(1)\n \n        # shared MLP\n        self.mlp = nn.Sequential(\n            # Conv2d比Linear方便操作\n            # nn.Linear(channel, channel // reduction, bias=False)\n            nn.Conv2d(channel, channel // reduction, 1, bias=False),\n            # inplace=True直接替换，节省内存\n            nn.ReLU(inplace=True),\n            # nn.Linear(channel // reduction, channel,bias=False)\n            nn.Conv2d(channel // reduction, channel, 1, bias=False)\n        )\n \n        # spatial attention\n        self.conv = nn.Conv2d(2, 1, kernel_size=spatial_kernel,\n                              padding=spatial_kernel // 2, bias=False)\n        self.sigmoid = nn.Sigmoid()\n \n    def forward(self, x):\n        max_out = self.mlp(self.max_pool(x))\n        avg_out = self.mlp(self.avg_pool(x))\n        channel_out = self.sigmoid(max_out + avg_out)\n        x = channel_out * x\n \n        max_out, _ = torch.max(x, dim=1, keepdim=True)\n        avg_out = torch.mean(x, dim=1, keepdim=True)\n        spatial_out = self.sigmoid(self.conv(torch.cat([max_out, avg_out], dim=1)))\n        x = spatial_out * x\n        return x\n\n\n\nclass GeM(nn.Module):\n    def __init__(self, p_init=3.0, eps=1e-6):\n        super().__init__()\n        self.p = nn.Parameter(torch.tensor(p_init, dtype=torch.float32))\n        self.eps = eps\n\n    def forward(self, x):\n        p = self.p.clamp(min=1e-6)\n        x = x.clamp(min=self.eps)\n        return x.pow(p).mean(dim=(-2, -1)).pow(1./p)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-19T11:12:53.194295Z","iopub.execute_input":"2025-04-19T11:12:53.194571Z","iopub.status.idle":"2025-04-19T11:12:53.202814Z","shell.execute_reply.started":"2025-04-19T11:12:53.194542Z","shell.execute_reply":"2025-04-19T11:12:53.201991Z"},"jupyter":{"source_hidden":true}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class CFG:\n    seed = 42\n    apex = False\n    print_freq = 100\n    num_workers = 2\n\n    OUTPUT_DIR = '/kaggle/working/'\n\n    train_datadir = '/kaggle/input/birdclef-2025/train_audio'\n    train_csv = '/kaggle/input/birdclef-2025/train.csv'\n    test_soundscapes = '/kaggle/input/birdclef-2025/test_soundscapes'\n    submission_csv = '/kaggle/input/birdclef-2025/sample_submission.csv'\n    taxonomy_csv = '/kaggle/input/birdclef-2025/taxonomy.csv'\n\n    spectrogram_npy = '/kaggle/input/birdclef25-mel-spectrograms/birdclef2025_melspec_5sec_256_256.npy'\n\n    model_name = 'efficientnet_b0'\n    pretrained = True\n    in_channels = 1\n\n    LOAD_DATA = True\n    FS = 32000\n    TARGET_DURATION = 5.0\n    TARGET_SHAPE = (256, 256)\n\n    N_FFT = 1024\n    HOP_LENGTH = 512\n    N_MELS = 128\n    FMIN = 50\n    FMAX = 14000\n\n    device = 'cuda' if torch.cuda.is_available() else 'cpu'\n\n    criterion = 'BCEWithLogitsLoss'\n\n    n_fold = 5\n    selected_folds = [0, 1, 2, 3, 4]\n\n    optimizer = 'AdamW'\n    lr = 5e-4\n    weight_decay = 1e-5\n\n    scheduler = 'CosineAnnealingLR'\n    min_lr = 1e-6\n\n    # 以下为需要调整的参数\n    #debug模式为 1Fold 2epochs 小批量训练\n    debug = False\n\n    # 模型参数，其余参数为自学习\n    epochs = 10\n    T_max = epochs #无需调整\n    batch_size = 32\n\n    # mixup启用\n    mixup_alpha = 0.5  # cfg.mixup_alpha = 0.5 启用mixup, =-1 关闭\n\n    # 频谱增强启用\n    time_masking = True\n    frequency_masking = True\n\n    def update_debug_settings(self):\n        if self.debug:\n            self.epochs = 2\n            self.selected_folds = [0]\n\n\ncfg = CFG()","metadata":{"trusted":true,"jupyter":{"source_hidden":true},"execution":{"iopub.status.busy":"2025-04-19T11:12:53.203602Z","iopub.execute_input":"2025-04-19T11:12:53.203895Z","iopub.status.idle":"2025-04-19T11:12:53.323556Z","shell.execute_reply.started":"2025-04-19T11:12:53.203869Z","shell.execute_reply":"2025-04-19T11:12:53.322790Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# 需要调整的参数\n#debug模式为 1Fold 2epochs 小批量训练\ncfg.debug = True\n\n# 模型参数，其余参数为自学习\ncfg.epochs = 10#图\ncfg.batch_size = 32#表\n\n#消融表\n# mixup启用\ncfg.mixup_alpha = 0.5  # cfg.mixup_alpha = 0.5 启用mixup, =-1 关闭\n\n# 频谱增强启用\ncfg.time_masking = True\ncfg.frequency_masking = True","metadata":{"trusted":true,"jupyter":{"source_hidden":true},"execution":{"iopub.status.busy":"2025-04-19T11:12:53.324982Z","iopub.execute_input":"2025-04-19T11:12:53.325214Z","iopub.status.idle":"2025-04-19T11:12:53.362950Z","shell.execute_reply.started":"2025-04-19T11:12:53.325197Z","shell.execute_reply":"2025-04-19T11:12:53.362332Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def plot_training_metrics(metrics, save_path=None):\n    \"\"\"\n    绘制训练过程中的指标变化\n    \n    Args:\n        metrics (dict): 包含训练和验证指标的字典\n        save_path (str, optional): 保存图像的路径，如果为None则显示图像\n    \"\"\"\n    plt.figure(figsize=(15, 10))\n    \n    # 创建两个子图\n    plt.subplot(2, 1, 1)\n    plt.plot(range(1, len(metrics['train_aucs']) + 1), metrics['train_aucs'], 'b-', label='Training AUC')\n    plt.plot(range(1, len(metrics['val_aucs']) + 1), metrics['val_aucs'], 'r-', label='Validation AUC')\n    plt.axvline(x=metrics['best_epoch'], color='g', linestyle='--', label=f'Best Epoch: {metrics[\"best_epoch\"]}')\n    plt.grid(True)\n    plt.xlabel('Epoch')\n    plt.ylabel('AUC Score')\n    plt.title('AUC vs. Epoch')\n    plt.legend()\n    \n    plt.subplot(2, 1, 2)\n    plt.plot(range(1, len(metrics['train_losses']) + 1), metrics['train_losses'], 'b-', label='Training Loss')\n    plt.plot(range(1, len(metrics['val_losses']) + 1), metrics['val_losses'], 'r-', label='Validation Loss')\n    plt.axvline(x=metrics['best_epoch'], color='g', linestyle='--', label=f'Best Epoch: {metrics[\"best_epoch\"]}')\n    plt.grid(True)\n    plt.xlabel('Epoch')\n    plt.ylabel('Loss')\n    plt.title('Loss vs. Epoch')\n    plt.legend()\n    \n    plt.tight_layout()\n    \n    if save_path:\n        plt.savefig(save_path)\n        print(f\"Training metrics plot saved to {save_path}\")\n    else:\n        plt.show()","metadata":{"trusted":true,"jupyter":{"source_hidden":true},"execution":{"iopub.status.busy":"2025-04-19T11:12:53.363579Z","iopub.execute_input":"2025-04-19T11:12:53.363740Z","iopub.status.idle":"2025-04-19T11:12:53.380614Z","shell.execute_reply.started":"2025-04-19T11:12:53.363726Z","shell.execute_reply":"2025-04-19T11:12:53.379939Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"#main\n\ndef set_seed(seed=42):\n    \"\"\"\n    Set seed for reproducibility\n    \"\"\"\n    random.seed(seed)\n    os.environ[\"PYTHONHASHSEED\"] = str(seed)\n    np.random.seed(seed)\n    torch.manual_seed(seed)\n    torch.cuda.manual_seed(seed)\n    torch.cuda.manual_seed_all(seed)\n    torch.backends.cudnn.deterministic = True\n    torch.backends.cudnn.benchmark = False\n\n\nset_seed(cfg.seed)\n\n\ndef audio2melspec(audio_data, cfg):\n    \"\"\"Convert audio data to mel spectrogram\"\"\"\n    if np.isnan(audio_data).any():\n        mean_signal = np.nanmean(audio_data)\n        audio_data = np.nan_to_num(audio_data, nan=mean_signal)\n\n    mel_spec = librosa.feature.melspectrogram(\n        y=audio_data,\n        sr=cfg.FS,\n        n_fft=cfg.N_FFT,\n        hop_length=cfg.HOP_LENGTH,\n        n_mels=cfg.N_MELS,\n        fmin=cfg.FMIN,\n        fmax=cfg.FMAX,\n        power=2.0\n    )\n\n    mel_spec_db = librosa.power_to_db(mel_spec, ref=np.max)\n    mel_spec_norm = (mel_spec_db - mel_spec_db.min()) / (mel_spec_db.max() - mel_spec_db.min() + 1e-8)\n\n    return mel_spec_norm\n\n\ndef process_audio_file(audio_path, cfg):\n    \"\"\"Process a single audio file to get the mel spectrogram\"\"\"\n    try:\n        audio_data, _ = librosa.load(audio_path, sr=cfg.FS)\n\n        target_samples = int(cfg.TARGET_DURATION * cfg.FS)\n\n        if len(audio_data) < target_samples:\n            n_copy = math.ceil(target_samples / len(audio_data))\n            if n_copy > 1:\n                audio_data = np.concatenate([audio_data] * n_copy)\n\n        # Extract center 5 seconds\n        start_idx = max(0, int(len(audio_data) / 2 - target_samples / 2))\n        end_idx = min(len(audio_data), start_idx + target_samples)\n        center_audio = audio_data[start_idx:end_idx]\n\n        if len(center_audio) < target_samples:\n            center_audio = np.pad(center_audio,\n                                  (0, target_samples - len(center_audio)),\n                                  mode='constant')\n\n        mel_spec = audio2melspec(center_audio, cfg)\n\n        if mel_spec.shape != cfg.TARGET_SHAPE:\n            mel_spec = cv2.resize(mel_spec, cfg.TARGET_SHAPE, interpolation=cv2.INTER_LINEAR)\n\n        return mel_spec.astype(np.float32)\n\n    except Exception as e:\n        print(f\"Error processing {audio_path}: {e}\")\n        return None\n\n\ndef generate_spectrograms(df, cfg):\n    \"\"\"Generate spectrograms from audio files\"\"\"\n    print(\"Generating mel spectrograms from audio files...\")\n    start_time = time.time()\n\n    all_bird_data = {}\n    errors = []\n\n    for i, row in tqdm(df.iterrows(), total=len(df)):\n        if cfg.debug and i >= 1000:\n            break\n\n        try:\n            samplename = row['samplename']\n            filepath = row['filepath']\n\n            mel_spec = process_audio_file(filepath, cfg)\n\n            if mel_spec is not None:\n                all_bird_data[samplename] = mel_spec\n\n        except Exception as e:\n            print(f\"Error processing {row.filepath}: {e}\")\n            errors.append((row.filepath, str(e)))\n\n    end_time = time.time()\n    print(f\"Processing completed in {end_time - start_time:.2f} seconds\")\n    print(f\"Successfully processed {len(all_bird_data)} files out of {len(df)}\")\n    print(f\"Failed to process {len(errors)} files\")\n\n    return all_bird_data\n\n\nclass BirdCLEFDatasetFromNPY(Dataset):\n    def __init__(self, df, cfg, spectrograms=None, mode=\"train\"):\n        self.df = df\n        self.cfg = cfg\n        self.mode = mode\n\n        self.spectrograms = spectrograms\n\n        taxonomy_df = pd.read_csv(self.cfg.taxonomy_csv)\n        self.species_ids = taxonomy_df['primary_label'].tolist()\n        self.num_classes = len(self.species_ids)\n        self.label_to_idx = {label: idx for idx, label in enumerate(self.species_ids)}\n\n        if 'filepath' not in self.df.columns:\n            self.df['filepath'] = self.cfg.train_datadir + '/' + self.df.filename\n\n        if 'samplename' not in self.df.columns:\n            self.df['samplename'] = self.df.filename.map(\n                lambda x: x.split('/')[0] + '-' + x.split('/')[-1].split('.')[0])\n\n        sample_names = set(self.df['samplename'])\n        if self.spectrograms:\n            found_samples = sum(1 for name in sample_names if name in self.spectrograms)\n            print(f\"Found {found_samples} matching spectrograms for {mode} dataset out of {len(self.df)} samples\")\n\n        if cfg.debug:\n            self.df = self.df.sample(min(1000, len(self.df)), random_state=cfg.seed).reset_index(drop=True)\n\n    def __len__(self):\n        return len(self.df)\n\n    def __getitem__(self, idx):\n        row = self.df.iloc[idx]\n        samplename = row['samplename']\n        spec = None\n\n        if self.spectrograms and samplename in self.spectrograms:\n            spec = self.spectrograms[samplename]\n        elif not self.cfg.LOAD_DATA:\n            spec = process_audio_file(row['filepath'], self.cfg)\n\n        if spec is None:\n            spec = np.zeros(self.cfg.TARGET_SHAPE, dtype=np.float32)\n            if self.mode == \"train\":  # Only print warning during training\n                print(f\"Warning: Spectrogram for {samplename} not found and could not be generated\")\n\n        spec = torch.tensor(spec, dtype=torch.float32).unsqueeze(0)  # Add channel dimension\n\n        if self.mode == \"train\":\n            spec = self.apply_spec_augmentations(spec)\n\n        target = self.encode_label(row['primary_label'])\n\n        if 'secondary_labels' in row and row['secondary_labels'] not in [[''], None, np.nan]:\n            if isinstance(row['secondary_labels'], str):\n                secondary_labels = eval(row['secondary_labels'])\n            else:\n                secondary_labels = row['secondary_labels']\n\n            for label in secondary_labels:\n                if label in self.label_to_idx:\n                    target[self.label_to_idx[label]] = 1.0\n\n        return {\n            'melspec': spec,\n            'target': torch.tensor(target, dtype=torch.float32),\n            'filename': row['filename']\n        }\n\n    def apply_spec_augmentations(self, spec):\n        \"\"\"Apply augmentations to spectrogram\"\"\"\n\n        # Time masking (horizontal stripes)\n        if random.random() < 0.3 and cfg.time_masking:\n            num_masks = random.randint(1, 3)\n            for _ in range(num_masks):\n                width = random.randint(5, 20)\n                start = random.randint(0, spec.shape[2] - width)\n                spec[0, :, start:start + width] = 0\n\n        # Frequency masking (vertical stripes)\n        if random.random() < 0.3 and cfg.frequency_masking:\n            num_masks = random.randint(1, 3)\n            for _ in range(num_masks):\n                height = random.randint(5, 20)\n                start = random.randint(0, spec.shape[1] - height)\n                spec[0, start:start + height, :] = 0\n\n        # Random brightness/contrast\n        '''if random.random() < 0.3:\n            gain = random.uniform(0.8, 1.2)\n            bias = random.uniform(-0.1, 0.1)\n            spec = spec * gain + bias\n            spec = torch.clamp(spec, 0, 1)'''\n\n        return spec\n\n    def apply_audio_augmentations(self, audio):\n        \"\"\"\n         对音频进行数据增强\n         参数:\n             audio (np.ndarray): 音频波形数据\n             cfg: 配置对象\n\n         返回:\n             np.ndarray: 增强后的音频波形\n         \"\"\"\n        # 复制音频以防止修改原始数据\n        augmented_audio = audio.copy()\n\n        # 1. 时间伸缩 (Time stretching)\n        if np.random.rand() < 0.5:\n            stretch_rate = np.random.uniform(0.8, 1.2)\n            augmented_audio = librosa.effects.time_stretch(augmented_audio, rate=stretch_rate)\n\n            # 确保音频长度一致\n            if len(augmented_audio) > cfg.DURATION * cfg.FS:\n                augmented_audio = augmented_audio[:cfg.DURATION * cfg.FS]\n            elif len(augmented_audio) < cfg.DURATION * cfg.FS:\n                # 填充不足的部分\n                pad_len = cfg.DURATION * cfg.FS - len(augmented_audio)\n                augmented_audio = np.pad(augmented_audio, (0, pad_len), 'constant')\n\n        # 2. 音高偏移 (Pitch shifting)\n        if np.random.rand() < 0.5:\n            n_steps = np.random.randint(-3, 4)\n            augmented_audio = librosa.effects.pitch_shift(augmented_audio, sr=cfg.FS, n_steps=n_steps)\n\n        # 3. 添加背景噪声\n        if np.random.rand() < 0.3:\n            noise_factor = np.random.uniform(0.001, 0.02)\n            noise = np.random.randn(len(augmented_audio))\n            augmented_audio = augmented_audio + noise_factor * noise\n\n        # 4. 音量变化\n        if np.random.rand() < 0.5:\n            gain = np.random.uniform(0.8, 1.2)\n            augmented_audio = augmented_audio * gain\n\n        # 剪裁幅度，确保在[-1, 1]范围内\n        augmented_audio = np.clip(augmented_audio, -1, 1)\n\n        return augmented_audio\n\n    def encode_label(self, label):\n        \"\"\"Encode label to one-hot vector\"\"\"\n        target = np.zeros(self.num_classes)\n        if label in self.label_to_idx:\n            target[self.label_to_idx[label]] = 1.0\n        return target\n\n\ndef collate_fn(batch):\n    \"\"\"Custom collate function to handle different sized spectrograms\"\"\"\n    batch = [item for item in batch if item is not None]\n    if len(batch) == 0:\n        return {}\n\n    result = {key: [] for key in batch[0].keys()}\n\n    for item in batch:\n        for key, value in item.items():\n            result[key].append(value)\n\n    for key in result:\n        if key == 'target' and isinstance(result[key][0], torch.Tensor):\n            result[key] = torch.stack(result[key])\n        elif key == 'melspec' and isinstance(result[key][0], torch.Tensor):\n            shapes = [t.shape for t in result[key]]\n            if len(set(str(s) for s in shapes)) == 1:\n                result[key] = torch.stack(result[key])\n\n    return result\n\n\nclass BirdCLEFModel(nn.Module):\n    def __init__(self, cfg):\n        super().__init__()\n        self.cfg = cfg\n\n        taxonomy_df = pd.read_csv(cfg.taxonomy_csv)\n        cfg.num_classes = len(taxonomy_df)\n\n        self.backbone = timm.create_model(\n            cfg.model_name,\n            pretrained=cfg.pretrained,\n            in_chans=cfg.in_channels,\n            drop_rate=0.2,\n            drop_path_rate=0.2\n        )\n\n        if 'efficientnet' in cfg.model_name:\n            backbone_out = self.backbone.classifier.in_features\n            self.backbone.classifier = nn.Identity()\n        elif 'resnet' in cfg.model_name:\n            backbone_out = self.backbone.fc.in_features\n            self.backbone.fc = nn.Identity()\n        else:\n            backbone_out = self.backbone.get_classifier().in_features\n            self.backbone.reset_classifier(0, '')\n\n        self.cbam = CBAM(backbone_out)\n        self.pool = GeM()\n        # self.pooling = nn.AdaptiveAvgPool2d(1)\n\n        self.feat_dim = backbone_out\n\n        # self.classifier = nn.Linear(backbone_out, cfg.num_classes)\n        # 分类头\n        self.classifier = nn.Sequential(\n            nn.Linear(backbone_out, 512),\n            nn.BatchNorm1d(512),\n            nn.SiLU(),\n            nn.Dropout(0.3),\n            nn.Linear(512, cfg.num_classes)\n        )\n\n        self.mixup_enabled = hasattr(cfg, 'mixup_alpha') and cfg.mixup_alpha > 0\n        if self.mixup_enabled:\n            self.mixup_alpha = cfg.mixup_alpha\n\n    def forward(self, x, targets):\n\n        if self.training and self.mixup_enabled and targets is not None:\n            mixed_x, targets_a, targets_b, lam = self.mixup_data(x, targets)\n            x = mixed_x\n        else:\n            targets_a, targets_b, lam = None, None, None\n\n        features = self.backbone(x)\n\n        if isinstance(features, dict):\n            features = features['features']\n\n        if len(features.shape) == 4:\n            features = self.cbam(features)\n            features = self.pool(features)\n            features = features.view(features.size(0), -1)\n\n        logits = self.classifier(features)\n\n        if self.training and self.mixup_enabled and targets is not None:\n            loss = self.mixup_criterion(F.binary_cross_entropy_with_logits,\n                                        logits, targets_a, targets_b, lam)\n            return logits, loss\n\n        return logits\n\n    def mixup_data(self, x, targets):\n        \"\"\"Applies mixup to the data batch\"\"\"\n        batch_size = x.size(0)\n\n        lam = np.random.beta(self.mixup_alpha, self.mixup_alpha)\n\n        indices = torch.randperm(batch_size).to(x.device)\n\n        mixed_x = lam * x + (1 - lam) * x[indices]\n\n        return mixed_x, targets, targets[indices], lam\n\n    def mixup_criterion(self, criterion, pred, y_a, y_b, lam):\n        \"\"\"Applies mixup to the loss function\"\"\"\n        return lam * criterion(pred, y_a) + (1 - lam) * criterion(pred, y_b)\n\n\ndef get_optimizer(model, cfg):\n    if cfg.optimizer == 'Adam':\n        optimizer = optim.Adam(\n            model.parameters(),\n            lr=cfg.lr,\n            weight_decay=cfg.weight_decay\n        )\n    elif cfg.optimizer == 'AdamW':\n        optimizer = optim.AdamW(\n            model.parameters(),\n            lr=cfg.lr,\n            weight_decay=cfg.weight_decay\n        )\n    elif cfg.optimizer == 'SGD':\n        optimizer = optim.SGD(\n            model.parameters(),\n            lr=cfg.lr,\n            momentum=0.9,\n            weight_decay=cfg.weight_decay\n        )\n    else:\n        raise NotImplementedError(f\"Optimizer {cfg.optimizer} not implemented\")\n\n    return optimizer\n\n\ndef get_scheduler(optimizer, cfg):\n    if cfg.scheduler == 'CosineAnnealingLR':\n        scheduler = lr_scheduler.CosineAnnealingLR(\n            optimizer,\n            T_max=cfg.T_max,\n            eta_min=cfg.min_lr\n        )\n    elif cfg.scheduler == 'ReduceLROnPlateau':\n        scheduler = lr_scheduler.ReduceLROnPlateau(\n            optimizer,\n            mode='min',\n            factor=0.5,\n            patience=2,\n            min_lr=cfg.min_lr,\n            verbose=True\n        )\n    elif cfg.scheduler == 'StepLR':\n        scheduler = lr_scheduler.StepLR(\n            optimizer,\n            step_size=cfg.epochs // 3,\n            gamma=0.5\n        )\n    elif cfg.scheduler == 'OneCycleLR':\n        scheduler = None\n    else:\n        scheduler = None\n\n    return scheduler\n\n\ndef get_criterion(cfg):\n    if cfg.criterion == 'BCEWithLogitsLoss':\n        criterion = nn.BCEWithLogitsLoss()\n    else:\n        raise NotImplementedError(f\"Criterion {cfg.criterion} not implemented\")\n\n    return criterion\n\n\ndef train_one_epoch(model, loader, optimizer, criterion, device, scheduler=None):\n    model.train()\n    losses = []\n    all_targets = []\n    all_outputs = []\n\n    pbar = tqdm(enumerate(loader), total=len(loader), desc=\"Training\")\n\n    for step, batch in pbar:\n\n        if isinstance(batch['melspec'], list):\n            batch_outputs = []\n            batch_losses = []\n\n            for i in range(len(batch['melspec'])):\n                inputs = batch['melspec'][i].unsqueeze(0).to(device)\n                target = batch['target'][i].unsqueeze(0).to(device)\n\n                optimizer.zero_grad()\n                output = model(inputs, target)\n                loss = criterion(output, target)\n                loss.backward()\n\n                batch_outputs.append(output.detach().cpu())\n                batch_losses.append(loss.item())\n\n            optimizer.step()\n            outputs = torch.cat(batch_outputs, dim=0).numpy()\n            loss = np.mean(batch_losses)\n            targets = batch['target'].numpy()\n\n        else:\n            inputs = batch['melspec'].to(device)\n            targets = batch['target'].to(device)\n\n            optimizer.zero_grad()\n            outputs = model(inputs,targets)\n\n            if isinstance(outputs, tuple):\n                outputs, loss = outputs\n            else:\n                loss = criterion(outputs, targets)\n\n            loss.backward()\n            optimizer.step()\n\n            outputs = outputs.detach().cpu().numpy()\n            targets = targets.detach().cpu().numpy()\n\n        if scheduler is not None and isinstance(scheduler, lr_scheduler.OneCycleLR):\n            scheduler.step()\n\n        all_outputs.append(outputs)\n        all_targets.append(targets)\n        losses.append(loss if isinstance(loss, float) else loss.item())\n\n        pbar.set_postfix({\n            'train_loss': np.mean(losses[-10:]) if losses else 0,\n            'lr': optimizer.param_groups[0]['lr']\n        })\n\n    all_outputs = np.concatenate(all_outputs)\n    all_targets = np.concatenate(all_targets)\n    auc = calculate_auc(all_targets, all_outputs)\n    avg_loss = np.mean(losses)\n\n    return avg_loss, auc\n\n\ndef validate(model, loader, criterion, device):\n    model.eval()\n    losses = []\n    all_targets = []\n    all_outputs = []\n\n    with torch.no_grad():\n        for batch in tqdm(loader, desc=\"Validation\"):\n            if isinstance(batch['melspec'], list):\n                batch_outputs = []\n                batch_losses = []\n\n                for i in range(len(batch['melspec'])):\n                    inputs = batch['melspec'][i].unsqueeze(0).to(device)\n                    target = batch['target'][i].unsqueeze(0).to(device)\n\n                    output = model(inputs, target)\n                    loss = criterion(output, target)\n\n                    batch_outputs.append(output.detach().cpu())\n                    batch_losses.append(loss.item())\n\n                outputs = torch.cat(batch_outputs, dim=0).numpy()\n                loss = np.mean(batch_losses)\n                targets = batch['target'].numpy()\n\n            else:\n                inputs = batch['melspec'].to(device)\n                targets = batch['target'].to(device)\n\n                outputs = model(inputs, targets)\n                loss = criterion(outputs, targets)\n\n                outputs = outputs.detach().cpu().numpy()\n                targets = targets.detach().cpu().numpy()\n\n            all_outputs.append(outputs)\n            all_targets.append(targets)\n            losses.append(loss if isinstance(loss, float) else loss.item())\n\n    all_outputs = np.concatenate(all_outputs)\n    all_targets = np.concatenate(all_targets)\n\n    auc = calculate_auc(all_targets, all_outputs)\n    avg_loss = np.mean(losses)\n\n    return avg_loss, auc\n\n\ndef calculate_auc(targets, outputs):\n    num_classes = targets.shape[1]\n    aucs = []\n\n    probs = 1 / (1 + np.exp(-outputs))\n\n    for i in range(num_classes):\n\n        if np.sum(targets[:, i]) > 0:\n            class_auc = roc_auc_score(targets[:, i], probs[:, i])\n            aucs.append(class_auc)\n\n    return np.mean(aucs) if aucs else 0.0\n\n\ndef run_training(df, cfg):\n    \"\"\"Training function that uses a simple train-validation split instead of cross-validation\"\"\"\n\n    taxonomy_df = pd.read_csv(cfg.taxonomy_csv)\n    species_ids = taxonomy_df['primary_label'].tolist()\n    cfg.num_classes = len(species_ids)\n\n    if cfg.debug:\n        cfg.update_debug_settings()\n\n    spectrograms = None\n    if cfg.LOAD_DATA:\n        print(\"Loading pre-computed mel spectrograms from NPY file...\")\n        try:\n            spectrograms = np.load(cfg.spectrogram_npy, allow_pickle=True).item()\n            print(f\"Loaded {len(spectrograms)} pre-computed mel spectrograms\")\n        except Exception as e:\n            print(f\"Error loading pre-computed spectrograms: {e}\")\n            print(\"Will generate spectrograms on-the-fly instead.\")\n            cfg.LOAD_DATA = False\n\n    if not cfg.LOAD_DATA:\n        print(\"Will generate spectrograms on-the-fly during training.\")\n        if 'filepath' not in df.columns:\n            df['filepath'] = cfg.train_datadir + '/' + df.filename\n        if 'samplename' not in df.columns:\n            df['samplename'] = df.filename.map(lambda x: x.split('/')[0] + '-' + x.split('/')[-1].split('.')[0])\n\n    # 使用简单的训练/验证分割代替k折交叉验证\n    # 按照物种标签进行分层抽样，保持训练集和验证集中的类别分布相似\n    from sklearn.model_selection import train_test_split\n    train_df, val_df = train_test_split(\n        df, \n        test_size=0.2, \n        random_state=cfg.seed, \n        stratify=df['primary_label']\n    )\n    \n    train_df = train_df.reset_index(drop=True)\n    val_df = val_df.reset_index(drop=True)\n\n    print(f'Training set: {len(train_df)} samples')\n    print(f'Validation set: {len(val_df)} samples')\n\n    # 创建训练和验证数据集\n    train_dataset = BirdCLEFDatasetFromNPY(train_df, cfg, spectrograms=spectrograms, mode='train')\n    val_dataset = BirdCLEFDatasetFromNPY(val_df, cfg, spectrograms=spectrograms, mode='valid')\n\n    train_loader = DataLoader(\n        train_dataset,\n        batch_size=cfg.batch_size,\n        shuffle=True,\n        num_workers=cfg.num_workers,\n        pin_memory=True,\n        collate_fn=collate_fn,\n        drop_last=True\n    )\n\n    val_loader = DataLoader(\n        val_dataset,\n        batch_size=cfg.batch_size,\n        shuffle=False,\n        num_workers=cfg.num_workers,\n        pin_memory=True,\n        collate_fn=collate_fn\n    )\n\n    # 创建模型和优化器\n    model = BirdCLEFModel(cfg).to(cfg.device)\n    optimizer = get_optimizer(model, cfg)\n    criterion = get_criterion(cfg)\n\n    if cfg.scheduler == 'OneCycleLR':\n        scheduler = lr_scheduler.OneCycleLR(\n            optimizer,\n            max_lr=cfg.lr,\n            steps_per_epoch=len(train_loader),\n            epochs=cfg.epochs,\n            pct_start=0.1\n        )\n    else:\n        scheduler = get_scheduler(optimizer, cfg)\n\n    best_auc = 0\n    best_epoch = 0\n    \n    # 用于跟踪训练和验证指标\n    train_losses = []\n    train_aucs = []\n    val_losses = []\n    val_aucs = []\n\n    for epoch in range(cfg.epochs):\n        print(f\"\\nEpoch {epoch + 1}/{cfg.epochs}\")\n\n        train_loss, train_auc = train_one_epoch(\n            model,\n            train_loader,\n            optimizer,\n            criterion,\n            cfg.device,\n            scheduler if isinstance(scheduler, lr_scheduler.OneCycleLR) else None\n        )\n\n        val_loss, val_auc = validate(model, val_loader, criterion, cfg.device)\n\n        # 记录指标用于后续绘图\n        train_losses.append(train_loss)\n        train_aucs.append(train_auc)\n        val_losses.append(val_loss)\n        val_aucs.append(val_auc)\n\n        if scheduler is not None and not isinstance(scheduler, lr_scheduler.OneCycleLR):\n            if isinstance(scheduler, lr_scheduler.ReduceLROnPlateau):\n                scheduler.step(val_loss)\n            else:\n                scheduler.step()\n\n        print(f\"Train Loss: {train_loss:.4f}, Train AUC: {train_auc:.4f}\")\n        print(f\"Val Loss: {val_loss:.4f}, Val AUC: {val_auc:.4f}\")\n\n        if val_auc > best_auc:\n            best_auc = val_auc\n            best_epoch = epoch + 1\n            print(f\"New best AUC: {best_auc:.4f} at epoch {best_epoch}\")\n\n            torch.save({\n                'model_state_dict': model.state_dict(),\n                'optimizer_state_dict': optimizer.state_dict(),\n                'scheduler_state_dict': scheduler.state_dict() if scheduler else None,\n                'epoch': epoch,\n                'val_auc': val_auc,\n                'train_auc': train_auc,\n                'cfg': cfg\n            }, \"best_model.pth\")\n\n    print(\"\\n\" + \"=\" * 60)\n    print(\"Training Results:\")\n    print(f\"Best AUC: {best_auc:.4f} at epoch {best_epoch}\")\n    print(\"=\" * 60)\n    \n    # 返回训练过程中的指标，用于后续绘图\n    return {\n        'train_losses': train_losses,\n        'train_aucs': train_aucs,\n        'val_losses': val_losses,\n        'val_aucs': val_aucs,\n        'best_auc': best_auc,\n        'best_epoch': best_epoch\n    }\n\n\nif __name__ == \"__main__\":\n    import time\n\n    print(\"\\nLoading training data...\")\n    train_df = pd.read_csv(cfg.train_csv)\n    taxonomy_df = pd.read_csv(cfg.taxonomy_csv)\n\n    print(\"\\nStarting training...\")\n    print(f\"LOAD_DATA is set to {cfg.LOAD_DATA}\")\n    if cfg.LOAD_DATA:\n        print(\"Using pre-computed mel spectrograms from NPY file\")\n    else:\n        print(\"Will generate spectrograms on-the-fly during training\")\n\n    # 运行训练并获取训练指标\n    metrics = run_training(train_df, cfg)\n    \n    # 绘制训练指标图表\n    plot_training_metrics(metrics, save_path=os.path.join(cfg.OUTPUT_DIR, \"training_metrics.png\"))\n\n    print(\"\\nTraining complete!\")\n","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"execution":{"iopub.status.busy":"2025-04-19T11:13:04.156269Z","iopub.execute_input":"2025-04-19T11:13:04.156569Z","iopub.status.idle":"2025-04-19T11:14:11.671307Z","shell.execute_reply.started":"2025-04-19T11:13:04.156543Z","shell.execute_reply":"2025-04-19T11:14:11.670146Z"},"jupyter":{"source_hidden":true}},"outputs":[],"execution_count":null}]}