{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":19596,"databundleVersionId":1292430,"sourceType":"competition"},{"sourceId":21669,"databundleVersionId":1692278,"sourceType":"competition"},{"sourceId":91844,"databundleVersionId":11361821,"sourceType":"competition"},{"sourceId":11252438,"sourceType":"datasetVersion","datasetId":7031717}],"dockerImageVersionId":30918,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# **BirdCLEF 2025 Inference Notebook**\nThis notebook runs inference on BirdCLEF 2025 test soundscapes and generates a submission file. It supports both single model inference and ensemble inference with multiple models. You can find the pre-processing and training processes in the following notebooks:\n\n- [Transforming Audio-to-Mel Spec. | BirdCLEF'25](https://www.kaggle.com/code/kadircandrisolu/transforming-audio-to-mel-spec-birdclef-25)  \n- [EfficientNet B0 Pytorch [Train] | BirdCLEF'25](https://www.kaggle.com/code/kadircandrisolu/efficientnet-b0-pytorch-train-birdclef-25)\n\n**Features**\n- Audio Preprocessing\n- Test-Time Augmentation (TTA)","metadata":{}},{"cell_type":"code","source":"import os\nimport gc\nimport warnings\nimport logging\nimport time\nimport math\nimport cv2\nfrom pathlib import Path\n\nimport numpy as np\nimport pandas as pd\nimport librosa\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nimport timm\nfrom tqdm.auto import tqdm\n\nwarnings.filterwarnings(\"ignore\")\nlogging.basicConfig(level=logging.ERROR)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:31:20.042935Z","iopub.execute_input":"2025-04-23T15:31:20.0432Z","iopub.status.idle":"2025-04-23T15:31:30.393283Z","shell.execute_reply.started":"2025-04-23T15:31:20.043169Z","shell.execute_reply":"2025-04-23T15:31:30.392588Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class CFG:\n \n    test_soundscapes = '/kaggle/input/birdclef-2025/train_soundscapes'\n    submission_csv = '/kaggle/input/birdclef-2025/sample_submission.csv'\n    taxonomy_csv = '/kaggle/input/birdclef-2025/taxonomy.csv'\n    model_path = '/kaggle/input/bird-clef-2025-eficientnetv2-b0-02-02-10epoch'  \n    \n    # Audio parameters\n    FS = 32000  \n    WINDOW_SIZE = 5  \n    \n    # Mel spectrogram parameters\n    N_FFT = 1024\n    HOP_LENGTH = 512\n    N_MELS = 128\n    FMIN = 50\n    FMAX = 14000\n    TARGET_SHAPE = (256, 256)\n    \n    model_name = 'tf_efficientnetv2_b0.in1k'\n    in_channels = 1\n    device = 'cpu'  \n    \n    # Inference parameters\n    batch_size = 16\n    use_tta = False  \n    tta_count = 3   \n    threshold = 0.5\n    \n    use_specific_folds = False  # If False, use all found models\n    folds = [1]  # Used only if use_specific_folds is True\n    \n    debug = False\n    debug_count = 3\n\ncfg = CFG()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:31:30.394023Z","iopub.execute_input":"2025-04-23T15:31:30.394223Z","iopub.status.idle":"2025-04-23T15:31:30.398945Z","shell.execute_reply.started":"2025-04-23T15:31:30.394205Z","shell.execute_reply":"2025-04-23T15:31:30.398065Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"print(f\"Using device: {cfg.device}\")\nprint(f\"Loading taxonomy data...\")\ntaxonomy_df = pd.read_csv(cfg.taxonomy_csv)\nspecies_ids = taxonomy_df['primary_label'].tolist()\nnum_classes = len(species_ids)\nprint(f\"Number of classes: {num_classes}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:31:30.399781Z","iopub.execute_input":"2025-04-23T15:31:30.400068Z","iopub.status.idle":"2025-04-23T15:31:30.429649Z","shell.execute_reply.started":"2025-04-23T15:31:30.400039Z","shell.execute_reply":"2025-04-23T15:31:30.428987Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class BirdCLEFModel(nn.Module):\n    def __init__(self, cfg, num_classes):\n        super().__init__()\n        self.cfg = cfg\n        \n        self.backbone = timm.create_model(\n            cfg.model_name,\n            pretrained=False,  \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.pooling = nn.AdaptiveAvgPool2d(1)\n        self.feat_dim = backbone_out\n        self.classifier = nn.Linear(backbone_out, num_classes)\n        \n    def forward(self, x):\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.pooling(features)\n            features = features.view(features.size(0), -1)\n        \n        logits = self.classifier(features)\n        return logits\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:31:30.431557Z","iopub.execute_input":"2025-04-23T15:31:30.431756Z","iopub.status.idle":"2025-04-23T15:31:30.438817Z","shell.execute_reply.started":"2025-04-23T15:31:30.431739Z","shell.execute_reply":"2025-04-23T15:31:30.438068Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def 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\ndef process_audio_segment(audio_data, cfg):\n    \"\"\"Process audio segment to get mel spectrogram\"\"\"\n    if len(audio_data) < cfg.FS * cfg.WINDOW_SIZE:\n        audio_data = np.pad(audio_data, \n                          (0, cfg.FS * cfg.WINDOW_SIZE - len(audio_data)), \n                          mode='constant')\n    \n    mel_spec = audio2melspec(audio_data, cfg)\n    \n    # Resize if needed\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)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:31:30.439815Z","iopub.execute_input":"2025-04-23T15:31:30.440036Z","iopub.status.idle":"2025-04-23T15:31:30.456933Z","shell.execute_reply.started":"2025-04-23T15:31:30.440006Z","shell.execute_reply":"2025-04-23T15:31:30.456261Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def find_model_files(cfg):\n    \"\"\"\n    Find all .pth model files in the specified model directory\n    \"\"\"\n    model_files = []\n    \n    model_dir = Path(cfg.model_path)\n    \n    for path in model_dir.glob('**/*.pth'):\n        model_files.append(str(path))\n    \n    return model_files\n\ndef load_models(cfg, num_classes):\n    \"\"\"\n    Load all found model files and prepare them for ensemble\n    \"\"\"\n    models = []\n    \n    model_files = find_model_files(cfg)\n    \n    if not model_files:\n        print(f\"Warning: No model files found under {cfg.model_path}!\")\n        return models\n    \n    print(f\"Found a total of {len(model_files)} model files.\")\n    \n    if cfg.use_specific_folds:\n        filtered_files = []\n        for fold in cfg.folds:\n            fold_files = [f for f in model_files if f\"fold{fold}\" in f]\n            filtered_files.extend(fold_files)\n        model_files = filtered_files\n        print(f\"Using {len(model_files)} model files for the specified folds ({cfg.folds}).\")\n    \n    for model_path in model_files:\n        try:\n            print(f\"Loading model: {model_path}\")\n            checkpoint = torch.load(model_path, map_location=torch.device(cfg.device))\n            \n            model = BirdCLEFModel(cfg, num_classes)\n            model.load_state_dict(checkpoint['model_state_dict'])\n            model = model.to(cfg.device)\n            model.eval()\n            \n            models.append(model)\n        except Exception as e:\n            print(f\"Error loading model {model_path}: {e}\")\n    \n    return models\n\ndef predict_on_spectrogram(audio_path, models, cfg, species_ids):\n    \"\"\"Process a single audio file and predict species presence for each 5-second segment\"\"\"\n    predictions = []\n    row_ids = []\n    soundscape_id = Path(audio_path).stem\n    \n    try:\n        print(f\"Processing {soundscape_id}\")\n        audio_data, _ = librosa.load(audio_path, sr=cfg.FS)\n        \n        total_segments = int(len(audio_data) / (cfg.FS * cfg.WINDOW_SIZE))\n        \n        for segment_idx in range(total_segments):\n            start_sample = segment_idx * cfg.FS * cfg.WINDOW_SIZE\n            end_sample = start_sample + cfg.FS * cfg.WINDOW_SIZE\n            segment_audio = audio_data[start_sample:end_sample]\n            \n            end_time_sec = (segment_idx + 1) * cfg.WINDOW_SIZE\n            row_id = f\"{soundscape_id}_{end_time_sec}\"\n            row_ids.append(row_id)\n\n            if cfg.use_tta:\n                tta_specs = apply_custom_tta(segment_audio, cfg.FS, tta_rounds=cfg.tta_count)\n                all_preds = []\n                for tta_spec in tta_specs:\n                    tta_tensor = torch.tensor(tta_spec, dtype=torch.float32).unsqueeze(0).to(cfg.device)  # [1, C, H, W]\n\n                    tta_tensor = tta_tensor.to(cfg.device)\n                    if len(models) == 1:\n                        with torch.no_grad():\n                            outputs = models[0](tta_tensor)\n                            probs = torch.sigmoid(outputs).cpu().numpy().squeeze()\n                            all_preds.append(probs)\n                    else:\n                        ensemble_preds = []\n                        for model in models:\n                            with torch.no_grad():\n                                outputs = model(tta_tensor)\n                                probs = torch.sigmoid(outputs).cpu().numpy().squeeze()\n                                ensemble_preds.append(probs)\n                        all_preds.append(np.mean(ensemble_preds, axis=0))\n            \n                final_preds = np.mean(all_preds, axis=0)\n            else:\n                mel_spec = process_audio_segment(segment_audio, cfg)\n                mel_spec = torch.tensor(mel_spec, dtype=torch.float32).unsqueeze(0).unsqueeze(0)\n                mel_spec = mel_spec.to(cfg.device)\n                \n                if len(models) == 1:\n                    with torch.no_grad():\n                        outputs = models[0](mel_spec)\n                        final_preds = torch.sigmoid(outputs).cpu().numpy().squeeze()\n                else:\n                    segment_preds = []\n                    for model in models:\n                        with torch.no_grad():\n                            outputs = model(mel_spec)\n                            probs = torch.sigmoid(outputs).cpu().numpy().squeeze()\n                            segment_preds.append(probs)\n\n                    final_preds = np.mean(segment_preds, axis=0)\n                    \n            predictions.append(final_preds)\n            \n    except Exception as e:\n        print(f\"Error processing {audio_path}: {e}\")\n    \n    return row_ids, predictions","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:31:30.457767Z","iopub.execute_input":"2025-04-23T15:31:30.458022Z","iopub.status.idle":"2025-04-23T15:31:30.477646Z","shell.execute_reply.started":"2025-04-23T15:31:30.458002Z","shell.execute_reply":"2025-04-23T15:31:30.476832Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def apply_tta(spec, tta_idx):\n    \"\"\"Apply test-time augmentation\"\"\"\n    if tta_idx == 0:\n        # Original spectrogram\n        return spec\n    elif tta_idx == 1:\n        # Time shift (horizontal flip)\n        return np.flip(spec, axis=1)\n    elif tta_idx == 2:\n        # Frequency shift (vertical flip)\n        return np.flip(spec, axis=0)\n    else:\n        return spec\n\ndef run_inference(cfg, models, species_ids):\n    \"\"\"Run inference on all test soundscapes\"\"\"\n    test_files = list(Path(cfg.test_soundscapes).glob('*.ogg'))\n    \n    if cfg.debug:\n        print(f\"Debug mode enabled, using only {cfg.debug_count} files\")\n        test_files = test_files[:cfg.debug_count]\n    \n    print(f\"Found {len(test_files)} test soundscapes\")\n\n    all_row_ids = []\n    all_predictions = []\n\n    for audio_path in tqdm(test_files):\n        row_ids, predictions = predict_on_spectrogram(str(audio_path), models, cfg, species_ids)\n        all_row_ids.extend(row_ids)\n        all_predictions.extend(predictions)\n    \n    return all_row_ids, all_predictions\n\ndef create_submission(row_ids, predictions, species_ids, cfg):\n    \"\"\"Create submission dataframe\"\"\"\n    print(\"Creating submission dataframe...\")\n\n    submission_dict = {'row_id': row_ids}\n    \n    for i, species in enumerate(species_ids):\n        submission_dict[species] = [pred[i] for pred in predictions]\n\n    submission_df = pd.DataFrame(submission_dict)\n\n    submission_df.set_index('row_id', inplace=True)\n\n    sample_sub = pd.read_csv(cfg.submission_csv, index_col='row_id')\n\n    missing_cols = set(sample_sub.columns) - set(submission_df.columns)\n    if missing_cols:\n        print(f\"Warning: Missing {len(missing_cols)} species columns in submission\")\n        for col in missing_cols:\n            submission_df[col] = 0.0\n\n    submission_df = submission_df[sample_sub.columns]\n\n    submission_df = submission_df.reset_index()\n    \n    return submission_df\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:31:30.478418Z","iopub.execute_input":"2025-04-23T15:31:30.478669Z","iopub.status.idle":"2025-04-23T15:31:30.492373Z","shell.execute_reply.started":"2025-04-23T15:31:30.478649Z","shell.execute_reply":"2025-04-23T15:31:30.491625Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"from albumentations import Compose\nfrom albumentations.core.transforms_interface import BasicTransform\n\ndef apply_custom_tta(audio_array, sr, tta_rounds=3):\n    \"\"\"基于 librosa 和 numpy 实现的 TTA，无需 albumentations\"\"\"\n    def random_crop(audio, sr, seconds=5):\n        length = int(sr * seconds)\n        if len(audio) < length:\n            pad = length - len(audio)\n            audio = np.pad(audio, (pad // 2, pad - pad // 2))\n            return audio\n        else:\n            max_start = len(audio) - length\n            if max_start <= 0:\n                return audio[:length]\n            start = np.random.randint(0, max_start)\n            return audio[start:start+length]\n    \n    def add_noise(audio, noise_level=0.02):\n        return audio + noise_level * np.random.randn(len(audio))\n\n    def to_mel(audio, sr):\n        mel = librosa.feature.melspectrogram(y=audio, sr=sr, n_mels=128, fmin=50, fmax=14000)\n        mel_db = librosa.power_to_db(mel)\n        return mel_db.astype(np.float32)\n\n    def spec_augment(spec):\n        spec = spec.copy()\n        freq_mask = np.random.randint(0, 20)\n        time_mask = np.random.randint(0, 30)\n        spec[:freq_mask, :] = 0\n        spec[:, :time_mask] = 0\n        return spec\n\n    def to_image(spec):\n        image = spec.astype(np.float32) / 100.0\n        return np.expand_dims(image, axis=0)  # [1, H, W]\n\n    specs = []\n    for _ in range(tta_rounds):\n        x = random_crop(audio_array, sr)\n        x = add_noise(x)\n        x = to_mel(x, sr)\n        x = spec_augment(x)\n        x = to_image(x)\n        specs.append(x)\n    \n    return specs","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:31:30.493167Z","iopub.execute_input":"2025-04-23T15:31:30.493363Z","iopub.status.idle":"2025-04-23T15:32:02.962678Z","shell.execute_reply.started":"2025-04-23T15:31:30.493346Z","shell.execute_reply":"2025-04-23T15:32:02.961985Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def main():\n    start_time = time.time()\n    print(\"Starting BirdCLEF-2025 inference...\")\n    print(f\"TTA enabled: {cfg.use_tta} (variations: {cfg.tta_count if cfg.use_tta else 0})\")\n\n    models = load_models(cfg, num_classes)\n    \n    if not models:\n        print(\"No models found! Please check model paths.\")\n        return\n    \n    print(f\"Model usage: {'Single model' if len(models) == 1 else f'Ensemble of {len(models)} models'}\")\n\n    row_ids, predictions = run_inference(cfg, models, species_ids)\n\n    submission_df = create_submission(row_ids, predictions, species_ids, cfg)\n\n    submission_path = 'submission.csv'\n    submission_df.to_csv(submission_path, index=False)\n    print(f\"Submission saved to {submission_path}\")\n    \n    end_time = time.time()\n    print(f\"Inference completed in {(end_time - start_time)/60:.2f} minutes\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:32:02.963688Z","iopub.execute_input":"2025-04-23T15:32:02.964238Z","iopub.status.idle":"2025-04-23T15:32:02.969147Z","shell.execute_reply.started":"2025-04-23T15:32:02.964207Z","shell.execute_reply":"2025-04-23T15:32:02.968461Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"if __name__ == \"__main__\":\n    cfg.use_tta = True         # 开启增强推理\n    cfg.tta_count = 3          # 每段音频增强3个版本\n    main()\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-04-23T15:32:02.969987Z","iopub.execute_input":"2025-04-23T15:32:02.970281Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}