Primero se carga la base de datos. Con esto se genera el archivo african-wildlife.yaml
In [1]:
yaml_content = """
path: african-wildlife
train: images/train
val: images/val
test: images/test
names:
0: buffalo
1: elephant
2: rhino
3: zebra
download: https://github.com/ultralytics/assets/releases/download/v0.0.0/african-wildlife.zip
"""
with open("african-wildlife.yaml", "w") as f:
f.write(yaml_content)
Análisis Exploratorio del dataset African Wildlife de Ultralytics. Primero se cambia el entorno de ejecución (runtime type) a t4gpu. El proceso necesita instalar ultralytics. El símbolo ! le dice al entorno que ejecute un comando de sistema (shell / terminal), no un comando de Python. El paquete también se intala en el entorno virtual.
In [2]:
!nvidia-smi
Sat Aug 23 01:17:27 2025
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 550.54.15 Driver Version: 550.54.15 CUDA Version: 12.4 |
|-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 Tesla T4 Off | 00000000:00:04.0 Off | 0 |
| N/A 42C P8 9W / 70W | 0MiB / 15360MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| No running processes found |
+-----------------------------------------------------------------------------------------+
In [3]:
!pip install ultralytics
Collecting ultralytics Downloading ultralytics-8.3.184-py3-none-any.whl.metadata (37 kB) Requirement already satisfied: numpy>=1.23.0 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (2.0.2) Requirement already satisfied: matplotlib>=3.3.0 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (3.10.0) Requirement already satisfied: opencv-python>=4.6.0 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (4.12.0.88) Requirement already satisfied: pillow>=7.1.2 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (11.3.0) Requirement already satisfied: pyyaml>=5.3.1 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (6.0.2) Requirement already satisfied: requests>=2.23.0 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (2.32.4) Requirement already satisfied: scipy>=1.4.1 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (1.16.1) Requirement already satisfied: torch>=1.8.0 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (2.8.0+cu126) Requirement already satisfied: torchvision>=0.9.0 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (0.23.0+cu126) Requirement already satisfied: tqdm>=4.64.0 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (4.67.1) Requirement already satisfied: psutil in /usr/local/lib/python3.12/dist-packages (from ultralytics) (5.9.5) Requirement already satisfied: py-cpuinfo in /usr/local/lib/python3.12/dist-packages (from ultralytics) (9.0.0) Requirement already satisfied: pandas>=1.1.4 in /usr/local/lib/python3.12/dist-packages (from ultralytics) (2.2.2) Collecting ultralytics-thop>=2.0.0 (from ultralytics) Downloading ultralytics_thop-2.0.16-py3-none-any.whl.metadata (14 kB) Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.12/dist-packages (from matplotlib>=3.3.0->ultralytics) (1.3.3) Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.12/dist-packages (from matplotlib>=3.3.0->ultralytics) (0.12.1) Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.12/dist-packages (from matplotlib>=3.3.0->ultralytics) (4.59.1) Requirement already satisfied: kiwisolver>=1.3.1 in /usr/local/lib/python3.12/dist-packages (from matplotlib>=3.3.0->ultralytics) (1.4.9) Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.12/dist-packages (from matplotlib>=3.3.0->ultralytics) (25.0) Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.12/dist-packages (from matplotlib>=3.3.0->ultralytics) (3.2.3) Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.12/dist-packages (from matplotlib>=3.3.0->ultralytics) (2.9.0.post0) Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.12/dist-packages (from pandas>=1.1.4->ultralytics) (2025.2) Requirement already satisfied: tzdata>=2022.7 in /usr/local/lib/python3.12/dist-packages (from pandas>=1.1.4->ultralytics) (2025.2) Requirement already satisfied: charset_normalizer<4,>=2 in /usr/local/lib/python3.12/dist-packages (from requests>=2.23.0->ultralytics) (3.4.3) Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.12/dist-packages (from requests>=2.23.0->ultralytics) (3.10) Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.12/dist-packages (from requests>=2.23.0->ultralytics) (2.5.0) Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.12/dist-packages (from requests>=2.23.0->ultralytics) (2025.8.3) Requirement already satisfied: filelock in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (3.19.1) Requirement already satisfied: typing-extensions>=4.10.0 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (4.14.1) Requirement already satisfied: setuptools in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (75.2.0) Requirement already satisfied: sympy>=1.13.3 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (1.13.3) Requirement already satisfied: networkx in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (3.5) Requirement already satisfied: jinja2 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (3.1.6) Requirement already satisfied: fsspec in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (2025.3.0) Requirement already satisfied: nvidia-cuda-nvrtc-cu12==12.6.77 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (12.6.77) Requirement already satisfied: nvidia-cuda-runtime-cu12==12.6.77 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (12.6.77) Requirement already satisfied: nvidia-cuda-cupti-cu12==12.6.80 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (12.6.80) Requirement already satisfied: nvidia-cudnn-cu12==9.10.2.21 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (9.10.2.21) Requirement already satisfied: nvidia-cublas-cu12==12.6.4.1 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (12.6.4.1) Requirement already satisfied: nvidia-cufft-cu12==11.3.0.4 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (11.3.0.4) Requirement already satisfied: nvidia-curand-cu12==10.3.7.77 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (10.3.7.77) Requirement already satisfied: nvidia-cusolver-cu12==11.7.1.2 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (11.7.1.2) Requirement already satisfied: nvidia-cusparse-cu12==12.5.4.2 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (12.5.4.2) Requirement already satisfied: nvidia-cusparselt-cu12==0.7.1 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (0.7.1) Requirement already satisfied: nvidia-nccl-cu12==2.27.3 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (2.27.3) Requirement already satisfied: nvidia-nvtx-cu12==12.6.77 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (12.6.77) Requirement already satisfied: nvidia-nvjitlink-cu12==12.6.85 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (12.6.85) Requirement already satisfied: nvidia-cufile-cu12==1.11.1.6 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (1.11.1.6) Requirement already satisfied: triton==3.4.0 in /usr/local/lib/python3.12/dist-packages (from torch>=1.8.0->ultralytics) (3.4.0) Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.12/dist-packages (from python-dateutil>=2.7->matplotlib>=3.3.0->ultralytics) (1.17.0) Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.12/dist-packages (from sympy>=1.13.3->torch>=1.8.0->ultralytics) (1.3.0) Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.12/dist-packages (from jinja2->torch>=1.8.0->ultralytics) (3.0.2) Downloading ultralytics-8.3.184-py3-none-any.whl (1.1 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.1/1.1 MB 24.1 MB/s eta 0:00:00 Downloading ultralytics_thop-2.0.16-py3-none-any.whl (28 kB) Installing collected packages: ultralytics-thop, ultralytics Successfully installed ultralytics-8.3.184 ultralytics-thop-2.0.16
Modelo YOLO¶
El siguiente comando entrena un modelo YOLO con el dataset African Wildlife. El paquete ultralytics se instala con el comando yolo (para entrenar el modelo) detect train indica que vamos a hacer detección de objetos y entrenar al modelo El data set .yaml incluye datos de entrenamiento y validación, así como las clases (especies animales). yolov8s.pt es un modelo preentrenado de yolo, y es mejor que yolo11n.pt el modelo analiza los datos en 30 ciclos. Las imágeness se redimensionan a 640×640 píxeles Batch define cuántas imágenes se usan en cada iteración.
El resultado del modelo genera las carpetas african-wildlife->images->(test,train,val), labels->(test,train,val) runs->detect->train->weights sample_data-> y los archivos yolo11n.pt y yolo8s.pt
In [4]:
# Start training from a pretrained *.pt model
!yolo detect train data=african-wildlife.yaml model=yolov8s.pt epochs=30 imgsz=640 batch=8
Creating new Ultralytics Settings v0.0.6 file ✅ View Ultralytics Settings with 'yolo settings' or at '/root/.config/Ultralytics/settings.json' Update Settings with 'yolo settings key=value', i.e. 'yolo settings runs_dir=path/to/dir'. For help see https://docs.ultralytics.com/quickstart/#ultralytics-settings. Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolov8s.pt to 'yolov8s.pt': 100% 21.5M/21.5M [00:00<00:00, 51.8MB/s] Ultralytics 8.3.184 🚀 Python-3.12.11 torch-2.8.0+cu126 CUDA:0 (Tesla T4, 15095MiB) engine/trainer: agnostic_nms=False, amp=True, augment=False, auto_augment=randaugment, batch=8, bgr=0.0, box=7.5, cache=False, cfg=None, classes=None, close_mosaic=10, cls=0.5, conf=None, copy_paste=0.0, copy_paste_mode=flip, cos_lr=False, cutmix=0.0, data=african-wildlife.yaml, degrees=0.0, deterministic=True, device=None, dfl=1.5, dnn=False, dropout=0.0, dynamic=False, embed=None, epochs=30, erasing=0.4, exist_ok=False, fliplr=0.5, flipud=0.0, format=torchscript, fraction=1.0, freeze=None, half=False, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, imgsz=640, int8=False, iou=0.7, keras=False, kobj=1.0, line_width=None, lr0=0.01, lrf=0.01, mask_ratio=4, max_det=300, mixup=0.0, mode=train, model=yolov8s.pt, momentum=0.937, mosaic=1.0, multi_scale=False, name=train, nbs=64, nms=False, opset=None, optimize=False, optimizer=auto, overlap_mask=True, patience=100, perspective=0.0, plots=True, pose=12.0, pretrained=True, profile=False, project=None, rect=False, resume=False, retina_masks=False, save=True, save_conf=False, save_crop=False, save_dir=runs/detect/train, save_frames=False, save_json=False, save_period=-1, save_txt=False, scale=0.5, seed=0, shear=0.0, show=False, show_boxes=True, show_conf=True, show_labels=True, simplify=True, single_cls=False, source=None, split=val, stream_buffer=False, task=detect, time=None, tracker=botsort.yaml, translate=0.1, val=True, verbose=True, vid_stride=1, visualize=False, warmup_bias_lr=0.1, warmup_epochs=3.0, warmup_momentum=0.8, weight_decay=0.0005, workers=8, workspace=None WARNING ⚠️ Dataset 'african-wildlife.yaml' images not found, missing path '/content/datasets/african-wildlife/images/val' Downloading https://ultralytics.com/assets/african-wildlife.zip to '/content/datasets/african-wildlife.zip': 100% 100M/100M [00:02<00:00, 46.4MB/s] Unzipping /content/datasets/african-wildlife.zip to /content/datasets/african-wildlife...: 100% 3018/3018 [00:00<00:00, 3999.86file/s] Dataset download success ✅ (3.7s), saved to /content/datasets Downloading https://ultralytics.com/assets/Arial.ttf to '/root/.config/Ultralytics/Arial.ttf': 100% 755k/755k [00:00<00:00, 21.2MB/s] Overriding model.yaml nc=80 with nc=4 from n params module arguments 0 -1 1 928 ultralytics.nn.modules.conv.Conv [3, 32, 3, 2] 1 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 2 -1 1 29056 ultralytics.nn.modules.block.C2f [64, 64, 1, True] 3 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2] 4 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True] 5 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 6 -1 2 788480 ultralytics.nn.modules.block.C2f [256, 256, 2, True] 7 -1 1 1180672 ultralytics.nn.modules.conv.Conv [256, 512, 3, 2] 8 -1 1 1838080 ultralytics.nn.modules.block.C2f [512, 512, 1, True] 9 -1 1 656896 ultralytics.nn.modules.block.SPPF [512, 512, 5] 10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 12 -1 1 591360 ultralytics.nn.modules.block.C2f [768, 256, 1] 13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 15 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1] 16 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 18 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1] 19 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2] 20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 21 -1 1 1969152 ultralytics.nn.modules.block.C2f [768, 512, 1] 22 [15, 18, 21] 1 2117596 ultralytics.nn.modules.head.Detect [4, [128, 256, 512]] Model summary: 129 layers, 11,137,148 parameters, 11,137,132 gradients, 28.7 GFLOPs Transferred 349/355 items from pretrained weights Freezing layer 'model.22.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt to 'yolo11n.pt': 100% 5.35M/5.35M [00:00<00:00, 65.4MB/s] AMP: checks passed ✅ train: Fast image access ✅ (ping: 0.0±0.0 ms, read: 1628.1±698.6 MB/s, size: 54.6 KB) train: Scanning /content/datasets/african-wildlife/labels/train... 1052 images, 0 backgrounds, 0 corrupt: 100% 1052/1052 [00:00<00:00, 2384.53it/s] train: New cache created: /content/datasets/african-wildlife/labels/train.cache albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, method='weighted_average', num_output_channels=3), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) val: Fast image access ✅ (ping: 0.0±0.0 ms, read: 451.1±225.8 MB/s, size: 42.2 KB) val: Scanning /content/datasets/african-wildlife/labels/val... 225 images, 0 backgrounds, 0 corrupt: 100% 225/225 [00:00<00:00, 2045.87it/s] val: New cache created: /content/datasets/african-wildlife/labels/val.cache Plotting labels to runs/detect/train/labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: AdamW(lr=0.00125, momentum=0.9) with parameter groups 57 weight(decay=0.0), 64 weight(decay=0.0005), 63 bias(decay=0.0) Image sizes 640 train, 640 val Using 2 dataloader workers Logging results to runs/detect/train Starting training for 30 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/30 1.88G 0.8357 1.618 1.22 9 640: 100% 132/132 [00:26<00:00, 4.90it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:04<00:00, 3.50it/s] all 225 379 0.8 0.751 0.829 0.632 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/30 2.63G 0.9346 1.184 1.264 16 640: 100% 132/132 [00:21<00:00, 6.04it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 6.37it/s] all 225 379 0.655 0.538 0.6 0.368 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/30 2.65G 1.016 1.215 1.315 18 640: 100% 132/132 [00:21<00:00, 6.09it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 7.71it/s] all 225 379 0.546 0.649 0.679 0.449 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/30 2.7G 1.049 1.266 1.356 12 640: 100% 132/132 [00:22<00:00, 5.99it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.61it/s] all 225 379 0.695 0.646 0.737 0.478 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/30 2.72G 1.015 1.165 1.316 21 640: 100% 132/132 [00:26<00:00, 5.07it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 6.05it/s] all 225 379 0.728 0.786 0.821 0.579 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/30 2.78G 0.9675 1.083 1.295 11 640: 100% 132/132 [00:21<00:00, 6.17it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 6.62it/s] all 225 379 0.827 0.731 0.84 0.597 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/30 2.8G 0.9589 1.044 1.286 13 640: 100% 132/132 [00:21<00:00, 6.07it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.49it/s] all 225 379 0.827 0.724 0.843 0.584 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/30 2.85G 0.918 0.9576 1.26 10 640: 100% 132/132 [00:22<00:00, 5.89it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.85it/s] all 225 379 0.9 0.811 0.89 0.655 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/30 2.87G 0.8834 0.953 1.241 50 640: 100% 132/132 [00:22<00:00, 5.91it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.50it/s] all 225 379 0.881 0.789 0.886 0.66 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/30 2.92G 0.8862 0.9173 1.232 27 640: 100% 132/132 [00:22<00:00, 5.88it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.46it/s] all 225 379 0.863 0.778 0.888 0.663 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/30 2.95G 0.8614 0.8738 1.207 15 640: 100% 132/132 [00:22<00:00, 5.83it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.43it/s] all 225 379 0.864 0.852 0.915 0.69 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/30 2.99G 0.8199 0.8105 1.186 18 640: 100% 132/132 [00:22<00:00, 5.81it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.75it/s] all 225 379 0.844 0.85 0.903 0.697 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/30 3.02G 0.795 0.7827 1.186 15 640: 100% 132/132 [00:22<00:00, 5.79it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.18it/s] all 225 379 0.897 0.83 0.914 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/30 3.07G 0.7838 0.7929 1.177 16 640: 100% 132/132 [00:22<00:00, 5.82it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.18it/s] all 225 379 0.859 0.83 0.916 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/30 3.09G 0.7604 0.7281 1.152 24 640: 100% 132/132 [00:22<00:00, 5.82it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 7.46it/s] all 225 379 0.896 0.844 0.915 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/30 3.14G 0.7542 0.7148 1.154 24 640: 100% 132/132 [00:22<00:00, 5.92it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 6.45it/s] all 225 379 0.894 0.87 0.923 0.731 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/30 3.16G 0.7455 0.705 1.139 18 640: 100% 132/132 [00:21<00:00, 6.13it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 6.46it/s] all 225 379 0.928 0.831 0.932 0.729 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/30 3.21G 0.7223 0.6931 1.129 12 640: 100% 132/132 [00:21<00:00, 6.09it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 7.27it/s] all 225 379 0.941 0.854 0.934 0.75 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/30 3.23G 0.6957 0.6472 1.108 14 640: 100% 132/132 [00:21<00:00, 6.01it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.61it/s] all 225 379 0.896 0.895 0.944 0.749 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/30 3.29G 0.6984 0.6313 1.11 15 640: 100% 132/132 [00:23<00:00, 5.69it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 7.60it/s] all 225 379 0.933 0.868 0.941 0.756 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, method='weighted_average', num_output_channels=3), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 21/30 3.31G 0.6153 0.5096 1.051 6 640: 100% 132/132 [00:24<00:00, 5.39it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.64it/s] all 225 379 0.922 0.87 0.942 0.758 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 22/30 3.36G 0.6005 0.489 1.044 6 640: 100% 132/132 [00:21<00:00, 6.04it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.66it/s] all 225 379 0.956 0.864 0.945 0.772 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 23/30 3.38G 0.584 0.4515 1.039 5 640: 100% 132/132 [00:22<00:00, 5.99it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.52it/s] all 225 379 0.93 0.895 0.955 0.787 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 24/30 3.43G 0.5627 0.4348 1.027 6 640: 100% 132/132 [00:21<00:00, 6.03it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.68it/s] all 225 379 0.948 0.887 0.957 0.793 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 25/30 3.62G 0.5428 0.4133 1.002 8 640: 100% 132/132 [00:21<00:00, 6.03it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.72it/s] all 225 379 0.934 0.891 0.952 0.794 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 26/30 3.67G 0.5261 0.3889 0.9929 9 640: 100% 132/132 [00:21<00:00, 6.11it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 6.99it/s] all 225 379 0.92 0.906 0.955 0.795 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 27/30 3.69G 0.5112 0.383 0.9826 6 640: 100% 132/132 [00:20<00:00, 6.30it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 6.72it/s] all 225 379 0.924 0.9 0.949 0.8 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 28/30 3.74G 0.4894 0.357 0.9697 10 640: 100% 132/132 [00:21<00:00, 6.26it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.65it/s] all 225 379 0.939 0.897 0.948 0.799 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 29/30 3.76G 0.4821 0.3559 0.9595 5 640: 100% 132/132 [00:22<00:00, 5.81it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.80it/s] all 225 379 0.943 0.895 0.954 0.806 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 30/30 3.81G 0.4708 0.335 0.959 5 640: 100% 132/132 [00:21<00:00, 6.04it/s] Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:01<00:00, 8.76it/s] all 225 379 0.941 0.912 0.957 0.815 30 epochs completed in 0.208 hours. Optimizer stripped from runs/detect/train/weights/last.pt, 22.5MB Optimizer stripped from runs/detect/train/weights/best.pt, 22.5MB Validating runs/detect/train/weights/best.pt... Ultralytics 8.3.184 🚀 Python-3.12.11 torch-2.8.0+cu126 CUDA:0 (Tesla T4, 15095MiB) Model summary (fused): 72 layers, 11,127,132 parameters, 0 gradients, 28.4 GFLOPs Class Images Instances Box(P R mAP50 mAP50-95): 100% 15/15 [00:02<00:00, 5.87it/s] all 225 379 0.941 0.912 0.957 0.816 buffalo 62 89 0.924 0.921 0.957 0.826 elephant 53 91 0.911 0.896 0.946 0.784 rhino 55 85 0.978 0.953 0.976 0.873 zebra 59 114 0.952 0.878 0.948 0.781 Speed: 0.2ms preprocess, 3.3ms inference, 0.0ms loss, 2.8ms postprocess per image Results saved to runs/detect/train 💡 Learn more at https://docs.ultralytics.com/modes/train
Análisis exploratorio de datos de acuerdo a https://www.kaggle.com/code/faldoae/exploratory-data-analysis-eda-for-image-datasets En este sitio las imágenes se organizan en carpetas diferentes a las de yolo, por lo que se debe adaptar el código.
In [5]:
import os
import yaml
from PIL import Image
import pandas as pd
# Leer YAML
with open('/content/datasets/african-wildlife/african-wildlife.yaml') as f:
data = yaml.safe_load(f)
# Carpetas base
base_img_dir = '/content/datasets/african-wildlife/images'
base_label_dir = '/content/datasets/african-wildlife/labels'
splits = ['train', 'val', 'test']
records = []
for split in splits:
img_dir = os.path.join(base_img_dir, split)
label_dir = os.path.join(base_label_dir, split)
if not os.path.exists(img_dir):
continue
for img_file in os.listdir(img_dir):
if img_file.lower().endswith(('.jpg', '.png', '.jpeg')):
img_path = os.path.join(img_dir, img_file)
img = Image.open(img_path)
w, h = img.size
label_path = os.path.join(label_dir, img_file.rsplit('.', 1)[0] + '.txt')
if os.path.exists(label_path):
with open(label_path, 'r') as lf:
lines = lf.readlines()
if lines:
first_class = int(lines[0].split()[0])
label_name = data['names'][first_class]
else:
label_name = 'no_label'
else:
label_name = 'no_label'
records.append({
'split': split,
'filepath': img_path,
'width': w,
'height': h,
'label': label_name
})
df = pd.DataFrame(records)
print(df.head(), "\nTotal imágenes:", len(df))
split filepath width height \
0 train /content/datasets/african-wildlife/images/trai... 500 332
1 train /content/datasets/african-wildlife/images/trai... 900 600
2 train /content/datasets/african-wildlife/images/trai... 1000 750
3 train /content/datasets/african-wildlife/images/trai... 768 500
4 train /content/datasets/african-wildlife/images/trai... 640 480
label
0 zebra
1 rhino
2 buffalo
3 elephant
4 zebra
Total imágenes: 1504
Estadísticas generales.
In [6]:
print("Total imágenes:", len(df))
print("\nDistribución por split:")
print(df['split'].value_counts())
print("\nDistribución por clase:")
print(df['label'].value_counts())
Total imágenes: 1504 Distribución por split: split train 1052 test 227 val 225 Name: count, dtype: int64 Distribución por clase: label zebra 380 buffalo 376 elephant 376 rhino 372 Name: count, dtype: int64
Conteo por especie en cada conjunto (split).
In [7]:
import seaborn as sns
import matplotlib.pyplot as plt
plt.figure(figsize=(10,5))
sns.countplot(data=df, x='label', hue='split', palette='Set2')
plt.title("Número de imágenes por especie y split")
plt.xticks(rotation=45)
plt.show()
Distribución de tamaños de imagen
In [8]:
plt.figure(figsize=(10,5))
sns.histplot(data=df, x='width', hue='split', bins=30, element='step')
plt.title("Distribución de ancho de imágenes por split")
plt.show()
plt.figure(figsize=(10,5))
sns.histplot(data=df, x='height', hue='split', bins=30, element='step')
plt.title("Distribución de alto de imágenes por split")
plt.show()
Relación ancho/alto por especie.
In [ ]:
df['ratio'] = df['width'] / df['height']
plt.figure(figsize=(10,5))
sns.boxplot(data=df, x='label', y='ratio')
plt.title("Relación ancho/alto por especie")
plt.xticks(rotation=45)
plt.show()
Mosaico de imágenes por especie.
In [ ]:
from PIL import Image
import random
def show_grid(df_subset, nrows=3, ncols=3):
fig, axes = plt.subplots(nrows, ncols, figsize=(ncols*3, nrows*3))
for ax, (_, row) in zip(axes.flatten(), df_subset.iterrows()):
img = Image.open(row['filepath'])
ax.imshow(img)
ax.set_title(row['label'])
ax.axis('off')
plt.tight_layout()
plt.show()
# Mostrar 9 imágenes aleatorias por clase
for label in df['label'].unique():
subset = df[df['label'] == label]
if len(subset) >= 9:
sample = subset.sample(9)
else:
sample = subset
print(f"Clase: {label}")
show_grid(sample)
Clase: buffalo
Clase: zebra
Clase: elephant
Clase: rhino
Distribución de clases.
In [ ]:
plt.figure(figsize=(8,8))
df['label'].value_counts().plot.pie(autopct='%1.1f%%', startangle=90, cmap='tab20')
plt.title("Proporción de clases en todo el dataset")
plt.ylabel("")
plt.show()
copiado de la ágina EDA
In [ ]:
#steps to convert image dataset to dataframe
#can use pathlib to search for files
#Here's an example of how to recursively search the current directory for files
#Then, you just need to reformat that into a dataframe. Here's how to do that
import pandas as pd
from pathlib import Path
data = '/content/datasets/african-wildlife/images/test'
paths = [path.parts[-2:] for path in
Path(data).rglob('*.*')] #writing purpose ('*.*') so that all image formats can be retrieved
df = pd.DataFrame(data=paths, columns=['Class','Images']) #create column names for dataframe
df = df.sort_values('Class',ascending=True) #sort class name
df.reset_index(drop=True, inplace=True) #sort index of each row
df #display dataframe
Out[ ]:
| Class | Images | |
|---|---|---|
| 0 | test | 2 (270).jpg |
| 1 | test | 1 (30).jpg |
| 2 | test | 2 (326).jpg |
| 3 | test | 1 (356).jpg |
| 4 | test | 3 (328).jpg |
| ... | ... | ... |
| 222 | test | 2 (341).jpg |
| 223 | test | 2 (85).jpg |
| 224 | test | 3 (123).jpg |
| 225 | test | 1 (192).jpg |
| 226 | test | 3 (320).jpg |
227 rows × 2 columns
In [ ]:
print('Count the number of image datasets')
print("Image Count : {}".format(len(df.Images)))
print("Class Count : {} \n".format(len(df['Class'].value_counts())))
print('Count the number of images in each class')
print(df['Class'].value_counts())
Count the number of image datasets Image Count : 227 Class Count : 1 Count the number of images in each class Class test 227 Name: count, dtype: int64
Adaptando el código de histogramas RGB
Para este histograma se utiliza el objeto df, el dataframe que contiene la información de las imágenes. Sin embargo, al correr el cósigo tuve un error por haber cambiado su valor en una celda anterior, así que tuve que volver a definirlo con df = pd.DataFrame(records)
In [ ]:
print(df.columns)
print(df.head())
Index(['Class', 'Images'], dtype='object') Class Images 0 test 4 (235).jpg 1 test 4 (252).jpg 2 test 3 (334).jpg 3 test 3 (314).jpg 4 test 3 (8).jpg
In [ ]:
for channel in ["R", "G", "B"]:
avg_df = df.groupby("Class").agg({
"Images": lambda paths: np.mean([calculate_channel_average(os.path.join(data, p), channel) for p in paths])
}).reset_index()
avg_df.rename(columns={"Images": f"{channel}_mean", "Class":"label"}, inplace=True)
channel_avgs.append(avg_df)
In [ ]:
records.append({
'split': split,
'filepath': img_path,
'width': w,
'height': h,
'label': label_name # 👈 aquí
})
df = pd.DataFrame(records)
El siguiente código nos genera una tabla con la composición de colores.
In [ ]:
import numpy as np
import cv2
def calculate_channel_average(img_path, channel):
"""
Calcular promedio de intensidades por canal (R, G o B).
"""
channel_dict = {"R": 2, "G": 1, "B": 0} # OpenCV usa BGR
channel_idx = channel_dict[channel]
img = cv2.imread(img_path) # carga como BGR
if img is None:
return np.nan
channel_intensities = img[:,:,channel_idx].flatten()
return np.mean(channel_intensities)
# Calcular promedios por clase
channel_avgs = []
for channel in ["R", "G", "B"]:
avg_df = df.groupby("label").agg({
"filepath": lambda paths: np.mean([calculate_channel_average(p, channel) for p in paths])
}).reset_index()
avg_df.rename(columns={"filepath": f"{channel}_mean"}, inplace=True)
channel_avgs.append(avg_df)
# Unir los tres dataframes
color_df = channel_avgs[0]
color_df["G_mean"] = channel_avgs[1]["G_mean"]
color_df["B_mean"] = channel_avgs[2]["B_mean"]
print(color_df)
label R_mean G_mean B_mean 0 buffalo 133.210377 127.855308 102.483883 1 elephant 129.002352 124.825104 106.761089 2 rhino 136.463383 128.335836 106.394742 3 zebra 132.278974 125.813713 100.741905
Histogramas de intensidad de colores por especie.
In [ ]:
def plot_channel_barplot(df, channel):
title_dict = {"R": "Red", "G": "Green", "B": "Blue"}
palette_dict = {"R": "Reds", "G": "Greens", "B": "Blues"}
plt.figure(figsize=(8,4))
values = df[f"{channel}_mean"].values
pal = sns.color_palette(palette_dict[channel], len(values))
rank = values.argsort().argsort()
ax = sns.barplot(x=df["label"], y=values, palette=np.array(pal[::-1])[rank])
plt.ylabel("Average Intensity")
plt.title(f"Average {title_dict[channel]} Channel Intensity per Class")
ax.set_xticklabels(ax.get_xticklabels(), rotation=45)
plt.show()
plot_channel_barplot(color_df, "R")
plot_channel_barplot(color_df, "G")
plot_channel_barplot(color_df, "B")
/tmp/ipython-input-78763786.py:10: FutureWarning: Passing `palette` without assigning `hue` is deprecated and will be removed in v0.14.0. Assign the `x` variable to `hue` and set `legend=False` for the same effect. ax = sns.barplot(x=df["label"], y=values, palette=np.array(pal[::-1])[rank]) /tmp/ipython-input-78763786.py:10: UserWarning: Numpy array is not a supported type for `palette`. Please convert your palette to a list. This will become an error in v0.14 ax = sns.barplot(x=df["label"], y=values, palette=np.array(pal[::-1])[rank]) /tmp/ipython-input-78763786.py:13: UserWarning: set_ticklabels() should only be used with a fixed number of ticks, i.e. after set_ticks() or using a FixedLocator. ax.set_xticklabels(ax.get_xticklabels(), rotation=45)
/tmp/ipython-input-78763786.py:10: FutureWarning: Passing `palette` without assigning `hue` is deprecated and will be removed in v0.14.0. Assign the `x` variable to `hue` and set `legend=False` for the same effect. ax = sns.barplot(x=df["label"], y=values, palette=np.array(pal[::-1])[rank]) /tmp/ipython-input-78763786.py:10: UserWarning: Numpy array is not a supported type for `palette`. Please convert your palette to a list. This will become an error in v0.14 ax = sns.barplot(x=df["label"], y=values, palette=np.array(pal[::-1])[rank]) /tmp/ipython-input-78763786.py:13: UserWarning: set_ticklabels() should only be used with a fixed number of ticks, i.e. after set_ticks() or using a FixedLocator. ax.set_xticklabels(ax.get_xticklabels(), rotation=45)
/tmp/ipython-input-78763786.py:10: FutureWarning: Passing `palette` without assigning `hue` is deprecated and will be removed in v0.14.0. Assign the `x` variable to `hue` and set `legend=False` for the same effect. ax = sns.barplot(x=df["label"], y=values, palette=np.array(pal[::-1])[rank]) /tmp/ipython-input-78763786.py:10: UserWarning: Numpy array is not a supported type for `palette`. Please convert your palette to a list. This will become an error in v0.14 ax = sns.barplot(x=df["label"], y=values, palette=np.array(pal[::-1])[rank]) /tmp/ipython-input-78763786.py:13: UserWarning: set_ticklabels() should only be used with a fixed number of ticks, i.e. after set_ticks() or using a FixedLocator. ax.set_xticklabels(ax.get_xticklabels(), rotation=45)
In [ ]:
color_df_melt = color_df.melt(id_vars="label", value_vars=["R_mean", "G_mean", "B_mean"],
var_name="Channel", value_name="Intensity")
plt.figure(figsize=(8,5))
sns.barplot(data=color_df_melt, x="label", y="Intensity", hue="Channel",
palette={"R_mean":"red", "G_mean":"green", "B_mean":"blue"})
plt.title("Average RGB Intensities per Class")
plt.xticks(rotation=45)
plt.show()
EDA de Github
FineTuning RetinaNet for Wildlife Detection with PyTorch¶
Al parecer los archivos descargados son parte de un entorno completo, y deben subirse en este punto la carpeta completa a colab. Veamos qué pasa. No supe si cargar el zip, terminé a subir los archivos manualmente.
El primer paso es instalar las dependencias de requirements.txt Me salió este error: WARNING: The following packages were previously imported in this runtime: [aiofiles,cv2,gradio] You must restart the runtime in order to use newly installed versions. me dió miedo y le dije que no.
In [10]:
%cd /content/project
!pip install -r requirements.txt
[Errno 2] No such file or directory: '/content/project' /content Collecting opencv-python==4.11.0.86 (from -r requirements.txt (line 1)) Downloading opencv_python-4.11.0.86-cp37-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (20 kB) Collecting torch==2.6.0 (from -r requirements.txt (line 2)) Downloading torch-2.6.0-cp312-cp312-manylinux1_x86_64.whl.metadata (28 kB) Collecting torchvision==0.21.0 (from -r requirements.txt (line 3)) Downloading torchvision-0.21.0-cp312-cp312-manylinux1_x86_64.whl.metadata (6.1 kB) Collecting torchaudio==2.6.0 (from -r requirements.txt (line 4)) Downloading torchaudio-2.6.0-cp312-cp312-manylinux1_x86_64.whl.metadata (6.6 kB) Collecting gradio==5.18.0 (from -r requirements.txt (line 5)) Downloading gradio-5.18.0-py3-none-any.whl.metadata (16 kB) Requirement already satisfied: numpy>=1.21.2 in /usr/local/lib/python3.12/dist-packages (from opencv-python==4.11.0.86->-r requirements.txt (line 1)) (2.0.2) Requirement already satisfied: filelock in /usr/local/lib/python3.12/dist-packages (from torch==2.6.0->-r requirements.txt (line 2)) (3.19.1) Requirement already satisfied: typing-extensions>=4.10.0 in /usr/local/lib/python3.12/dist-packages (from torch==2.6.0->-r requirements.txt (line 2)) (4.14.1) Requirement already satisfied: networkx in /usr/local/lib/python3.12/dist-packages (from torch==2.6.0->-r requirements.txt (line 2)) (3.5) Requirement already satisfied: jinja2 in /usr/local/lib/python3.12/dist-packages (from torch==2.6.0->-r requirements.txt (line 2)) (3.1.6) Requirement already satisfied: fsspec in /usr/local/lib/python3.12/dist-packages (from torch==2.6.0->-r requirements.txt (line 2)) (2025.3.0) Collecting nvidia-cuda-nvrtc-cu12==12.4.127 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_cuda_nvrtc_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl.metadata (1.5 kB) Collecting nvidia-cuda-runtime-cu12==12.4.127 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_cuda_runtime_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl.metadata (1.5 kB) Collecting nvidia-cuda-cupti-cu12==12.4.127 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_cuda_cupti_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl.metadata (1.6 kB) Collecting nvidia-cudnn-cu12==9.1.0.70 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_cudnn_cu12-9.1.0.70-py3-none-manylinux2014_x86_64.whl.metadata (1.6 kB) Collecting nvidia-cublas-cu12==12.4.5.8 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_cublas_cu12-12.4.5.8-py3-none-manylinux2014_x86_64.whl.metadata (1.5 kB) Collecting nvidia-cufft-cu12==11.2.1.3 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_cufft_cu12-11.2.1.3-py3-none-manylinux2014_x86_64.whl.metadata (1.5 kB) Collecting nvidia-curand-cu12==10.3.5.147 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_curand_cu12-10.3.5.147-py3-none-manylinux2014_x86_64.whl.metadata (1.5 kB) Collecting nvidia-cusolver-cu12==11.6.1.9 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_cusolver_cu12-11.6.1.9-py3-none-manylinux2014_x86_64.whl.metadata (1.6 kB) Collecting nvidia-cusparse-cu12==12.3.1.170 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_cusparse_cu12-12.3.1.170-py3-none-manylinux2014_x86_64.whl.metadata (1.6 kB) Collecting nvidia-cusparselt-cu12==0.6.2 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_cusparselt_cu12-0.6.2-py3-none-manylinux2014_x86_64.whl.metadata (6.8 kB) Collecting nvidia-nccl-cu12==2.21.5 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_nccl_cu12-2.21.5-py3-none-manylinux2014_x86_64.whl.metadata (1.8 kB) Collecting nvidia-nvtx-cu12==12.4.127 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_nvtx_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl.metadata (1.7 kB) Collecting nvidia-nvjitlink-cu12==12.4.127 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading nvidia_nvjitlink_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl.metadata (1.5 kB) Collecting triton==3.2.0 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading triton-3.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (1.4 kB) Requirement already satisfied: setuptools in /usr/local/lib/python3.12/dist-packages (from torch==2.6.0->-r requirements.txt (line 2)) (75.2.0) Collecting sympy==1.13.1 (from torch==2.6.0->-r requirements.txt (line 2)) Downloading sympy-1.13.1-py3-none-any.whl.metadata (12 kB) Requirement already satisfied: pillow!=8.3.*,>=5.3.0 in /usr/local/lib/python3.12/dist-packages (from torchvision==0.21.0->-r requirements.txt (line 3)) (11.3.0) Collecting aiofiles<24.0,>=22.0 (from gradio==5.18.0->-r requirements.txt (line 5)) Downloading aiofiles-23.2.1-py3-none-any.whl.metadata (9.7 kB) Requirement already satisfied: anyio<5.0,>=3.0 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (4.10.0) Requirement already satisfied: fastapi<1.0,>=0.115.2 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.116.1) Requirement already satisfied: ffmpy in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.6.1) Collecting gradio-client==1.7.2 (from gradio==5.18.0->-r requirements.txt (line 5)) Downloading gradio_client-1.7.2-py3-none-any.whl.metadata (7.1 kB) Requirement already satisfied: httpx>=0.24.1 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.28.1) Requirement already satisfied: huggingface-hub>=0.28.1 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.34.4) Collecting markupsafe~=2.0 (from gradio==5.18.0->-r requirements.txt (line 5)) Downloading MarkupSafe-2.1.5-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (3.0 kB) Requirement already satisfied: orjson~=3.0 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (3.11.2) Requirement already satisfied: packaging in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (25.0) Requirement already satisfied: pandas<3.0,>=1.0 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (2.2.2) Requirement already satisfied: pydantic>=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (2.11.7) Requirement already satisfied: pydub in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.25.1) Requirement already satisfied: python-multipart>=0.0.18 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.0.20) Requirement already satisfied: pyyaml<7.0,>=5.0 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (6.0.2) Requirement already satisfied: ruff>=0.9.3 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.12.9) Requirement already satisfied: safehttpx<0.2.0,>=0.1.6 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.1.6) Requirement already satisfied: semantic-version~=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (2.10.0) Requirement already satisfied: starlette<1.0,>=0.40.0 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.47.2) Requirement already satisfied: tomlkit<0.14.0,>=0.12.0 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.13.3) Requirement already satisfied: typer<1.0,>=0.12 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.16.0) Requirement already satisfied: uvicorn>=0.14.0 in /usr/local/lib/python3.12/dist-packages (from gradio==5.18.0->-r requirements.txt (line 5)) (0.35.0) Requirement already satisfied: websockets<16.0,>=10.0 in /usr/local/lib/python3.12/dist-packages (from gradio-client==1.7.2->gradio==5.18.0->-r requirements.txt (line 5)) (15.0.1) Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.12/dist-packages (from sympy==1.13.1->torch==2.6.0->-r requirements.txt (line 2)) (1.3.0) Requirement already satisfied: idna>=2.8 in /usr/local/lib/python3.12/dist-packages (from anyio<5.0,>=3.0->gradio==5.18.0->-r requirements.txt (line 5)) (3.10) Requirement already satisfied: sniffio>=1.1 in /usr/local/lib/python3.12/dist-packages (from anyio<5.0,>=3.0->gradio==5.18.0->-r requirements.txt (line 5)) (1.3.1) Requirement already satisfied: certifi in /usr/local/lib/python3.12/dist-packages (from httpx>=0.24.1->gradio==5.18.0->-r requirements.txt (line 5)) (2025.8.3) Requirement already satisfied: httpcore==1.* in /usr/local/lib/python3.12/dist-packages (from httpx>=0.24.1->gradio==5.18.0->-r requirements.txt (line 5)) (1.0.9) Requirement already satisfied: h11>=0.16 in /usr/local/lib/python3.12/dist-packages (from httpcore==1.*->httpx>=0.24.1->gradio==5.18.0->-r requirements.txt (line 5)) (0.16.0) Requirement already satisfied: requests in /usr/local/lib/python3.12/dist-packages (from huggingface-hub>=0.28.1->gradio==5.18.0->-r requirements.txt (line 5)) (2.32.4) Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.12/dist-packages (from huggingface-hub>=0.28.1->gradio==5.18.0->-r requirements.txt (line 5)) (4.67.1) Requirement already satisfied: hf-xet<2.0.0,>=1.1.3 in /usr/local/lib/python3.12/dist-packages (from huggingface-hub>=0.28.1->gradio==5.18.0->-r requirements.txt (line 5)) (1.1.7) Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio==5.18.0->-r requirements.txt (line 5)) (2.9.0.post0) Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio==5.18.0->-r requirements.txt (line 5)) (2025.2) Requirement already satisfied: tzdata>=2022.7 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio==5.18.0->-r requirements.txt (line 5)) (2025.2) Requirement already satisfied: annotated-types>=0.6.0 in /usr/local/lib/python3.12/dist-packages (from pydantic>=2.0->gradio==5.18.0->-r requirements.txt (line 5)) (0.7.0) Requirement already satisfied: pydantic-core==2.33.2 in /usr/local/lib/python3.12/dist-packages (from pydantic>=2.0->gradio==5.18.0->-r requirements.txt (line 5)) (2.33.2) Requirement already satisfied: typing-inspection>=0.4.0 in /usr/local/lib/python3.12/dist-packages (from pydantic>=2.0->gradio==5.18.0->-r requirements.txt (line 5)) (0.4.1) Requirement already satisfied: click>=8.0.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio==5.18.0->-r requirements.txt (line 5)) (8.2.1) Requirement already satisfied: shellingham>=1.3.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio==5.18.0->-r requirements.txt (line 5)) (1.5.4) Requirement already satisfied: rich>=10.11.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio==5.18.0->-r requirements.txt (line 5)) (13.9.4) Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.12/dist-packages (from python-dateutil>=2.8.2->pandas<3.0,>=1.0->gradio==5.18.0->-r requirements.txt (line 5)) (1.17.0) Requirement already satisfied: markdown-it-py>=2.2.0 in /usr/local/lib/python3.12/dist-packages (from rich>=10.11.0->typer<1.0,>=0.12->gradio==5.18.0->-r requirements.txt (line 5)) (4.0.0) Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /usr/local/lib/python3.12/dist-packages (from rich>=10.11.0->typer<1.0,>=0.12->gradio==5.18.0->-r requirements.txt (line 5)) (2.19.2) Requirement already satisfied: charset_normalizer<4,>=2 in /usr/local/lib/python3.12/dist-packages (from requests->huggingface-hub>=0.28.1->gradio==5.18.0->-r requirements.txt (line 5)) (3.4.3) Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.12/dist-packages (from requests->huggingface-hub>=0.28.1->gradio==5.18.0->-r requirements.txt (line 5)) (2.5.0) Requirement already satisfied: mdurl~=0.1 in /usr/local/lib/python3.12/dist-packages (from markdown-it-py>=2.2.0->rich>=10.11.0->typer<1.0,>=0.12->gradio==5.18.0->-r requirements.txt (line 5)) (0.1.2) Downloading opencv_python-4.11.0.86-cp37-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (63.0 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 63.0/63.0 MB 11.7 MB/s eta 0:00:00 Downloading torch-2.6.0-cp312-cp312-manylinux1_x86_64.whl (766.6 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 766.6/766.6 MB 1.9 MB/s eta 0:00:00 Downloading torchvision-0.21.0-cp312-cp312-manylinux1_x86_64.whl (7.2 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 7.2/7.2 MB 130.6 MB/s eta 0:00:00 Downloading torchaudio-2.6.0-cp312-cp312-manylinux1_x86_64.whl (3.4 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 3.4/3.4 MB 98.0 MB/s eta 0:00:00 Downloading gradio-5.18.0-py3-none-any.whl (62.3 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 62.3/62.3 MB 8.4 MB/s eta 0:00:00 Downloading gradio_client-1.7.2-py3-none-any.whl (322 kB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 322.1/322.1 kB 31.4 MB/s eta 0:00:00 Downloading nvidia_cublas_cu12-12.4.5.8-py3-none-manylinux2014_x86_64.whl (363.4 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 363.4/363.4 MB 1.5 MB/s eta 0:00:00 Downloading nvidia_cuda_cupti_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl (13.8 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 13.8/13.8 MB 93.5 MB/s eta 0:00:00 Downloading nvidia_cuda_nvrtc_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl (24.6 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 24.6/24.6 MB 84.2 MB/s eta 0:00:00 Downloading nvidia_cuda_runtime_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl (883 kB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 883.7/883.7 kB 63.6 MB/s eta 0:00:00 Downloading nvidia_cudnn_cu12-9.1.0.70-py3-none-manylinux2014_x86_64.whl (664.8 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 664.8/664.8 MB 2.9 MB/s eta 0:00:00 Downloading nvidia_cufft_cu12-11.2.1.3-py3-none-manylinux2014_x86_64.whl (211.5 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 211.5/211.5 MB 5.4 MB/s eta 0:00:00 Downloading nvidia_curand_cu12-10.3.5.147-py3-none-manylinux2014_x86_64.whl (56.3 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 56.3/56.3 MB 15.1 MB/s eta 0:00:00 Downloading nvidia_cusolver_cu12-11.6.1.9-py3-none-manylinux2014_x86_64.whl (127.9 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 127.9/127.9 MB 7.8 MB/s eta 0:00:00 Downloading nvidia_cusparse_cu12-12.3.1.170-py3-none-manylinux2014_x86_64.whl (207.5 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 207.5/207.5 MB 5.7 MB/s eta 0:00:00 Downloading nvidia_cusparselt_cu12-0.6.2-py3-none-manylinux2014_x86_64.whl (150.1 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 150.1/150.1 MB 6.7 MB/s eta 0:00:00 Downloading nvidia_nccl_cu12-2.21.5-py3-none-manylinux2014_x86_64.whl (188.7 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 188.7/188.7 MB 6.1 MB/s eta 0:00:00 Downloading nvidia_nvjitlink_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl (21.1 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 21.1/21.1 MB 108.5 MB/s eta 0:00:00 Downloading nvidia_nvtx_cu12-12.4.127-py3-none-manylinux2014_x86_64.whl (99 kB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 99.1/99.1 kB 10.0 MB/s eta 0:00:00 Downloading sympy-1.13.1-py3-none-any.whl (6.2 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 6.2/6.2 MB 130.1 MB/s eta 0:00:00 Downloading triton-3.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (253.2 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 253.2/253.2 MB 3.1 MB/s eta 0:00:00 Downloading aiofiles-23.2.1-py3-none-any.whl (15 kB) Downloading MarkupSafe-2.1.5-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (28 kB) Installing collected packages: triton, nvidia-cusparselt-cu12, sympy, opencv-python, nvidia-nvtx-cu12, nvidia-nvjitlink-cu12, nvidia-nccl-cu12, nvidia-curand-cu12, nvidia-cufft-cu12, nvidia-cuda-runtime-cu12, nvidia-cuda-nvrtc-cu12, nvidia-cuda-cupti-cu12, nvidia-cublas-cu12, markupsafe, aiofiles, nvidia-cusparse-cu12, nvidia-cudnn-cu12, nvidia-cusolver-cu12, gradio-client, torch, gradio, torchvision, torchaudio Attempting uninstall: triton Found existing installation: triton 3.4.0 Uninstalling triton-3.4.0: Successfully uninstalled triton-3.4.0 Attempting uninstall: nvidia-cusparselt-cu12 Found existing installation: nvidia-cusparselt-cu12 0.7.1 Uninstalling nvidia-cusparselt-cu12-0.7.1: Successfully uninstalled nvidia-cusparselt-cu12-0.7.1 Attempting uninstall: sympy Found existing installation: sympy 1.13.3 Uninstalling sympy-1.13.3: Successfully uninstalled sympy-1.13.3 Attempting uninstall: opencv-python Found existing installation: opencv-python 4.12.0.88 Uninstalling opencv-python-4.12.0.88: Successfully uninstalled opencv-python-4.12.0.88 Attempting uninstall: nvidia-nvtx-cu12 Found existing installation: nvidia-nvtx-cu12 12.6.77 Uninstalling nvidia-nvtx-cu12-12.6.77: Successfully uninstalled nvidia-nvtx-cu12-12.6.77 Attempting uninstall: nvidia-nvjitlink-cu12 Found existing installation: nvidia-nvjitlink-cu12 12.6.85 Uninstalling nvidia-nvjitlink-cu12-12.6.85: Successfully uninstalled nvidia-nvjitlink-cu12-12.6.85 Attempting uninstall: nvidia-nccl-cu12 Found existing installation: nvidia-nccl-cu12 2.27.3 Uninstalling nvidia-nccl-cu12-2.27.3: Successfully uninstalled nvidia-nccl-cu12-2.27.3 Attempting uninstall: nvidia-curand-cu12 Found existing installation: nvidia-curand-cu12 10.3.7.77 Uninstalling nvidia-curand-cu12-10.3.7.77: Successfully uninstalled nvidia-curand-cu12-10.3.7.77 Attempting uninstall: nvidia-cufft-cu12 Found existing installation: nvidia-cufft-cu12 11.3.0.4 Uninstalling nvidia-cufft-cu12-11.3.0.4: Successfully uninstalled nvidia-cufft-cu12-11.3.0.4 Attempting uninstall: nvidia-cuda-runtime-cu12 Found existing installation: nvidia-cuda-runtime-cu12 12.6.77 Uninstalling nvidia-cuda-runtime-cu12-12.6.77: Successfully uninstalled nvidia-cuda-runtime-cu12-12.6.77 Attempting uninstall: nvidia-cuda-nvrtc-cu12 Found existing installation: nvidia-cuda-nvrtc-cu12 12.6.77 Uninstalling nvidia-cuda-nvrtc-cu12-12.6.77: Successfully uninstalled nvidia-cuda-nvrtc-cu12-12.6.77 Attempting uninstall: nvidia-cuda-cupti-cu12 Found existing installation: nvidia-cuda-cupti-cu12 12.6.80 Uninstalling nvidia-cuda-cupti-cu12-12.6.80: Successfully uninstalled nvidia-cuda-cupti-cu12-12.6.80 Attempting uninstall: nvidia-cublas-cu12 Found existing installation: nvidia-cublas-cu12 12.6.4.1 Uninstalling nvidia-cublas-cu12-12.6.4.1: Successfully uninstalled nvidia-cublas-cu12-12.6.4.1 Attempting uninstall: markupsafe Found existing installation: MarkupSafe 3.0.2 Uninstalling MarkupSafe-3.0.2: Successfully uninstalled MarkupSafe-3.0.2 Attempting uninstall: aiofiles Found existing installation: aiofiles 24.1.0 Uninstalling aiofiles-24.1.0: Successfully uninstalled aiofiles-24.1.0 Attempting uninstall: nvidia-cusparse-cu12 Found existing installation: nvidia-cusparse-cu12 12.5.4.2 Uninstalling nvidia-cusparse-cu12-12.5.4.2: Successfully uninstalled nvidia-cusparse-cu12-12.5.4.2 Attempting uninstall: nvidia-cudnn-cu12 Found existing installation: nvidia-cudnn-cu12 9.10.2.21 Uninstalling nvidia-cudnn-cu12-9.10.2.21: Successfully uninstalled nvidia-cudnn-cu12-9.10.2.21 Attempting uninstall: nvidia-cusolver-cu12 Found existing installation: nvidia-cusolver-cu12 11.7.1.2 Uninstalling nvidia-cusolver-cu12-11.7.1.2: Successfully uninstalled nvidia-cusolver-cu12-11.7.1.2 Attempting uninstall: gradio-client Found existing installation: gradio_client 1.11.1 Uninstalling gradio_client-1.11.1: Successfully uninstalled gradio_client-1.11.1 Attempting uninstall: torch Found existing installation: torch 2.8.0+cu126 Uninstalling torch-2.8.0+cu126: Successfully uninstalled torch-2.8.0+cu126 Attempting uninstall: gradio Found existing installation: gradio 5.42.0 Uninstalling gradio-5.42.0: Successfully uninstalled gradio-5.42.0 Attempting uninstall: torchvision Found existing installation: torchvision 0.23.0+cu126 Uninstalling torchvision-0.23.0+cu126: Successfully uninstalled torchvision-0.23.0+cu126 Attempting uninstall: torchaudio Found existing installation: torchaudio 2.8.0+cu126 Uninstalling torchaudio-2.8.0+cu126: Successfully uninstalled torchaudio-2.8.0+cu126 Successfully installed aiofiles-23.2.1 gradio-5.18.0 gradio-client-1.7.2 markupsafe-2.1.5 nvidia-cublas-cu12-12.4.5.8 nvidia-cuda-cupti-cu12-12.4.127 nvidia-cuda-nvrtc-cu12-12.4.127 nvidia-cuda-runtime-cu12-12.4.127 nvidia-cudnn-cu12-9.1.0.70 nvidia-cufft-cu12-11.2.1.3 nvidia-curand-cu12-10.3.5.147 nvidia-cusolver-cu12-11.6.1.9 nvidia-cusparse-cu12-12.3.1.170 nvidia-cusparselt-cu12-0.6.2 nvidia-nccl-cu12-2.21.5 nvidia-nvjitlink-cu12-12.4.127 nvidia-nvtx-cu12-12.4.127 opencv-python-4.11.0.86 sympy-1.13.1 torch-2.6.0 torchaudio-2.6.0 torchvision-0.21.0 triton-3.2.0
In [11]:
from model import create_model
from config import NUM_CLASSES, DEVICE, CLASSES
from datasets import CustomDataset
from custom_utils import collate_fn
Los archivos .py cargados en el proyecto se ejecutan con %cd /content/ !python custom_utils.py Tuve que ajustar el primer archivo config.py para ajustarlo a los directorios donde guardé los datos. Al ejecutar train.py tuve tantos errores que terminé cambiando config.py muchas veces. También hubo que cambiar la ruta de los datos train y val.
In [12]:
%%writefile /content/config.py
# config.py
import torch
BATCH_SIZE = 8 # Increase / decrease according to GPU memeory.
RESIZE_TO = 640 # Resize the image for training and transforms.
NUM_EPOCHS = 60 # Number of epochs to train for.
NUM_WORKERS = 4 # Number of parallel workers for data loading.
DEVICE = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
# Training images and labels files directory.
TRAIN_DIR = "/content/datasets/african-wildlife/train"
# Validation images and labels files directory.
VALID_DIR = "/content/datasets/african-wildlife/val"
# Classes: 0 index is reserved for background.
CLASSES = ["__background__", "buffalo", "elephant", "rhino", "zebra"]
NUM_CLASSES = len(CLASSES)
# Whether to visualize images after crearing the data loaders.
VISUALIZE_TRANSFORMED_IMAGES = True
# Location to save model and plots.
OUT_DIR = "outputs"
Overwriting /content/config.py
El comando %cd /content/ !python custom_utils.py ejecuta los archivos .py, me dí cuenta que después ya no era necesario agregar %cd /content/ Al final este tampoco era necesario correrlo
In [ ]:
%cd /content/
!python custom_utils.py
/content
Al parecer no se ejecutan cada unos de los .py, porque algunos contienen instrucciones del código que se usarán en los .py que sí se ejecutan (Estos cargarán la información de los restantes), aunque el tutorial no dice qué archivos sí se ejecutan ¬¬.
In [ ]:
%cd /content/
!python model.py
/content
Downloading: "https://download.pytorch.org/models/retinanet_resnet50_fpn_v2_coco-5905b1c5.pth" to /root/.cache/torch/hub/checkpoints/retinanet_resnet50_fpn_v2_coco-5905b1c5.pth
100% 146M/146M [00:00<00:00, 185MB/s]
RetinaNet(
(backbone): BackboneWithFPN(
(body): IntermediateLayerGetter(
(conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
(layer1): Sequential(
(0): Bottleneck(
(conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer2): Sequential(
(0): Bottleneck(
(conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer3): Sequential(
(0): Bottleneck(
(conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(4): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(5): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer4): Sequential(
(0): Bottleneck(
(conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
)
(fpn): FeaturePyramidNetwork(
(inner_blocks): ModuleList(
(0): Conv2dNormActivation(
(0): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1))
)
(1): Conv2dNormActivation(
(0): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1))
)
(2): Conv2dNormActivation(
(0): Conv2d(2048, 256, kernel_size=(1, 1), stride=(1, 1))
)
)
(layer_blocks): ModuleList(
(0-2): 3 x Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
)
)
(extra_blocks): LastLevelP6P7(
(p6): Conv2d(2048, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
(p7): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
)
)
)
(anchor_generator): AnchorGenerator()
(head): RetinaNetHead(
(classification_head): RetinaNetClassificationHead(
(conv): Sequential(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
)
(cls_logits): Conv2d(256, 45, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
)
(regression_head): RetinaNetRegressionHead(
(conv): Sequential(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
)
(bbox_reg): Conv2d(256, 36, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
)
)
(transform): GeneralizedRCNNTransform(
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
Resize(min_size=(800,), max_size=1333, mode='bilinear')
)
)
36,414,865 total parameters.
36,189,521 training parameters.
Aquí necesito instalar torchmetrics para que no salga error.
In [13]:
!pip install torchmetrics
Collecting torchmetrics Downloading torchmetrics-1.8.1-py3-none-any.whl.metadata (22 kB) Requirement already satisfied: numpy>1.20.0 in /usr/local/lib/python3.12/dist-packages (from torchmetrics) (2.0.2) Requirement already satisfied: packaging>17.1 in /usr/local/lib/python3.12/dist-packages (from torchmetrics) (25.0) Requirement already satisfied: torch>=2.0.0 in /usr/local/lib/python3.12/dist-packages (from torchmetrics) (2.6.0) Collecting lightning-utilities>=0.8.0 (from torchmetrics) Downloading lightning_utilities-0.15.2-py3-none-any.whl.metadata (5.7 kB) Requirement already satisfied: setuptools in /usr/local/lib/python3.12/dist-packages (from lightning-utilities>=0.8.0->torchmetrics) (75.2.0) Requirement already satisfied: typing_extensions in /usr/local/lib/python3.12/dist-packages (from lightning-utilities>=0.8.0->torchmetrics) (4.14.1) Requirement already satisfied: filelock in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (3.19.1) Requirement already satisfied: networkx in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (3.5) Requirement already satisfied: jinja2 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (3.1.6) Requirement already satisfied: fsspec in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (2025.3.0) Requirement already satisfied: nvidia-cuda-nvrtc-cu12==12.4.127 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (12.4.127) Requirement already satisfied: nvidia-cuda-runtime-cu12==12.4.127 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (12.4.127) Requirement already satisfied: nvidia-cuda-cupti-cu12==12.4.127 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (12.4.127) Requirement already satisfied: nvidia-cudnn-cu12==9.1.0.70 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (9.1.0.70) Requirement already satisfied: nvidia-cublas-cu12==12.4.5.8 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (12.4.5.8) Requirement already satisfied: nvidia-cufft-cu12==11.2.1.3 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (11.2.1.3) Requirement already satisfied: nvidia-curand-cu12==10.3.5.147 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (10.3.5.147) Requirement already satisfied: nvidia-cusolver-cu12==11.6.1.9 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (11.6.1.9) Requirement already satisfied: nvidia-cusparse-cu12==12.3.1.170 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (12.3.1.170) Requirement already satisfied: nvidia-cusparselt-cu12==0.6.2 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (0.6.2) Requirement already satisfied: nvidia-nccl-cu12==2.21.5 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (2.21.5) Requirement already satisfied: nvidia-nvtx-cu12==12.4.127 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (12.4.127) Requirement already satisfied: nvidia-nvjitlink-cu12==12.4.127 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (12.4.127) Requirement already satisfied: triton==3.2.0 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (3.2.0) Requirement already satisfied: sympy==1.13.1 in /usr/local/lib/python3.12/dist-packages (from torch>=2.0.0->torchmetrics) (1.13.1) Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.12/dist-packages (from sympy==1.13.1->torch>=2.0.0->torchmetrics) (1.3.0) Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.12/dist-packages (from jinja2->torch>=2.0.0->torchmetrics) (2.1.5) Downloading torchmetrics-1.8.1-py3-none-any.whl (982 kB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 983.0/983.0 kB 25.2 MB/s eta 0:00:00 Downloading lightning_utilities-0.15.2-py3-none-any.whl (29 kB) Installing collected packages: lightning-utilities, torchmetrics Successfully installed lightning-utilities-0.15.2 torchmetrics-1.8.1
Train.py no encuentra las imágenes porque están en un orden diferente de carpetas. Hay que modificar datasets.py para que las busque bien.
In [14]:
%%writefile /content/datasets.py
# datasets.py
import torch
import cv2
import numpy as np
import os
import glob
from config import CLASSES, RESIZE_TO, TRAIN_DIR, BATCH_SIZE
from torch.utils.data import Dataset, DataLoader
from custom_utils import collate_fn, get_train_transform, get_valid_transform
class CustomDataset(Dataset):
def __init__(self, dir_path, width, height, classes, transforms=None):
"""
:param dir_path: Directory containing 'images/' and 'labels/' subfolders.
:param width: Resized image width.
:param height: Resized image height.
:param classes: List of class names (or an indexing scheme).
:param transforms: Albumentations transformations to apply.
"""
self.transforms = transforms
self.dir_path = dir_path
self.image_dir = os.path.join(self.dir_path, "images")
self.label_dir = os.path.join(self.dir_path, "labels")
self.width = width
self.height = height
self.classes = classes
# Gather all image paths
self.image_file_types = ["*.jpg", "*.jpeg", "*.png", "*.ppm", "*.JPG"]
self.all_image_paths = []
for file_type in self.image_file_types:
self.all_image_paths.extend(glob.glob(os.path.join(self.image_dir, file_type)))
# Sort for consistent ordering
self.all_image_paths = sorted(self.all_image_paths)
self.all_image_names = [os.path.basename(img_p) for img_p in self.all_image_paths]
def __len__(self):
return len(self.all_image_paths)
def __getitem__(self, idx):
# 1) Read image
image_name = self.all_image_names[idx]
image_path = os.path.join(self.image_dir, image_name)
label_filename = os.path.splitext(image_name)[0] + ".txt"
label_path = os.path.join(self.label_dir, label_filename)
image = cv2.imread(image_path)
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB).astype(np.float32)
# 2) Resize image (to the model's expected size)
image_resized = cv2.resize(image, (self.width, self.height))
image_resized /= 255.0 # Scale pixel values to [0, 1]
# 3) Read bounding boxes (normalized) from .txt file
boxes = []
labels = []
if os.path.exists(label_path):
with open(label_path, "r") as f:
lines = f.readlines()
for line in lines:
line = line.strip()
if not line:
continue
# Format: class_id x_min y_min x_max y_max (all in [0..1])
parts = line.split()
class_id = int(parts[0]) # e.g. 0, 1, 2, ...
xmin = float(parts[1])
ymin = float(parts[2])
xmax = float(parts[3])
ymax = float(parts[4])
# Example: if you want class IDs to start at 1 for foreground
# and background=0, do:
label_idx = class_id + 1
# Convert normalized coords to absolute (in resized space)
x_min_final = xmin * self.width
y_min_final = ymin * self.height
x_max_final = xmax * self.width
y_max_final = ymax * self.height
# Ensure valid box
if x_max_final <= x_min_final:
x_max_final = x_min_final + 1
if y_max_final <= y_min_final:
y_max_final = y_min_final + 1
# Clip if out of bounds
x_min_final = max(0, min(x_min_final, self.width - 1))
x_max_final = max(0, min(x_max_final, self.width))
y_min_final = max(0, min(y_min_final, self.height - 1))
y_max_final = max(0, min(y_max_final, self.height))
boxes.append([x_min_final, y_min_final, x_max_final, y_max_final])
labels.append(label_idx)
# 4) Convert boxes & labels to Torch tensors
if len(boxes) == 0:
boxes = torch.zeros((0, 4), dtype=torch.float32)
labels = torch.zeros((0,), dtype=torch.int64)
else:
boxes = torch.tensor(boxes, dtype=torch.float32)
labels = torch.tensor(labels, dtype=torch.int64)
# 5) Prepare the target dict
area = (
(boxes[:, 3] - boxes[:, 1]) * (boxes[:, 2] - boxes[:, 0])
if len(boxes) > 0
else torch.tensor([], dtype=torch.float32)
)
iscrowd = torch.zeros((len(boxes),), dtype=torch.int64)
image_id = torch.tensor([idx])
target = {"boxes": boxes, "labels": labels, "area": area, "iscrowd": iscrowd, "image_id": image_id}
# 6) Albumentations transforms: pass Python lists, not Tensors
if self.transforms:
bboxes_list = boxes.cpu().numpy().tolist() # shape: list of [xmin, ymin, xmax, ymax]
labels_list = labels.cpu().numpy().tolist() # shape: list of ints
transformed = self.transforms(
image=image_resized,
bboxes=bboxes_list,
labels=labels_list,
)
# Reassign the image
image_resized = transformed["image"]
# Convert bboxes back to Torch Tensors
new_bboxes_list = transformed["bboxes"] # list of [xmin, ymin, xmax, ymax]
new_labels_list = transformed["labels"] # list of int
if len(new_bboxes_list) > 0:
new_bboxes = torch.tensor(new_bboxes_list, dtype=torch.float32)
new_labels = torch.tensor(new_labels_list, dtype=torch.int64)
else:
new_bboxes = torch.zeros((0, 4), dtype=torch.float32)
new_labels = torch.zeros((0,), dtype=torch.int64)
target["boxes"] = new_bboxes
target["labels"] = new_labels
return image_resized, target
# ---------------------------------------------------------
# Create train/valid datasets and loaders
# ---------------------------------------------------------
def create_train_dataset(DIR):
train_dataset = CustomDataset(
dir_path=DIR, width=RESIZE_TO, height=RESIZE_TO, classes=CLASSES, transforms=get_train_transform()
)
return train_dataset
def create_valid_dataset(DIR):
valid_dataset = CustomDataset(
dir_path=DIR, width=RESIZE_TO, height=RESIZE_TO, classes=CLASSES, transforms=get_valid_transform()
)
return valid_dataset
def create_train_loader(train_dataset, num_workers=0):
train_loader = DataLoader(
train_dataset,
batch_size=BATCH_SIZE,
shuffle=True,
num_workers=num_workers,
collate_fn=collate_fn,
drop_last=True,
)
return train_loader
def create_valid_loader(valid_dataset, num_workers=0):
valid_loader = DataLoader(
valid_dataset,
batch_size=BATCH_SIZE,
shuffle=False,
num_workers=num_workers,
collate_fn=collate_fn,
drop_last=True,
)
return valid_loader
# ---------------------------------------------------------
# Debug/demo if run directly
# ---------------------------------------------------------
if __name__ == "__main__":
# Example usage with no transforms for debugging
dataset = CustomDataset(dir_path=TRAIN_DIR, width=RESIZE_TO, height=RESIZE_TO, classes=CLASSES, transforms=None)
print(f"Number of training images: {len(dataset)}")
def visualize_sample(image, target):
"""
Visualize a single sample using OpenCV. Expects
`image` as a NumPy array of shape (H, W, 3) in [0..1].
"""
# Convert [0,1] float -> [0,255] uint8
img = (image * 255).astype(np.uint8)
# Convert RGB -> BGR
img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
boxes = target["boxes"].cpu().numpy().astype(np.int32)
labels = target["labels"].cpu().numpy().astype(np.int32)
for i, box in enumerate(boxes):
x1, y1, x2, y2 = box
class_idx = labels[i]
# If your class_idx starts at 1 for "first class", ensure you handle that:
# e.g. if CLASSES = ["background", "class1", "class2", ...]
if 0 <= class_idx < len(CLASSES):
class_str = CLASSES[class_idx]
else:
class_str = f"Label_{class_idx}"
cv2.rectangle(img, (x1, y1), (x2, y2), (0, 0, 255), 2)
cv2.putText(img, class_str, (x1, max(y1 - 5, 0)), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 0, 255), 2)
cv2.imshow("Sample", img)
cv2.waitKey(0)
# Visualize a few samples
NUM_SAMPLES_TO_VISUALIZE = 10
for i in range(NUM_SAMPLES_TO_VISUALIZE):
image, target = dataset[i] # No transforms in this example
# `image` is shape (H, W, 3) in [0..1]
print(f"Visualizing sample {i}, boxes: {target['boxes'].shape[0]}")
visualize_sample(image, target)
cv2.destroyAllWindows()
Overwriting /content/datasets.py
Después de tanto intentar ajustar las rutas, me rendí y voy a tratar de acomodar todo en las carpetas como espera el código.
In [15]:
# Crear carpetas
!mkdir -p /content/datasets/african-wildlife/train/images
!mkdir -p /content/datasets/african-wildlife/train/labels
!mkdir -p /content/datasets/african-wildlife/val/images
!mkdir -p /content/datasets/african-wildlife/val/labels
!mkdir -p /content/datasets/african-wildlife/test/images
!mkdir -p /content/datasets/african-wildlife/test/labels
# Mover imágenes
!cp /content/datasets/african-wildlife/images/train/* /content/datasets/african-wildlife/train/images/
!cp /content/datasets/african-wildlife/images/val/* /content/datasets/african-wildlife/val/images/
!cp /content/datasets/african-wildlife/images/test/* /content/datasets/african-wildlife/test/images/
# Mover labels
!cp /content/datasets/african-wildlife/labels/train/* /content/datasets/african-wildlife/train/labels/
!cp /content/datasets/african-wildlife/labels/val/* /content/datasets/african-wildlife/val/labels/
!cp /content/datasets/african-wildlife/labels/test/* /content/datasets/african-wildlife/test/labels/
Lo que hice después de 1 trillón de intentos: cambiar las imágenes y etiquetas de carpeta con el código de arriba, luego cambiar la ruta en la que busca los datos en config.py con este ajuste:
In [16]:
%cd /content/
!python train.py
/content
/usr/local/lib/python3.12/dist-packages/albumentations/core/composition.py:331: UserWarning: Got processor for bboxes, but no transform to process it.
self._set_keys()
/usr/local/lib/python3.12/dist-packages/torch/utils/data/dataloader.py:624: UserWarning: This DataLoader will create 4 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.
warnings.warn(
Number of training samples: 1052
Number of validation samples: 225
Downloading: "https://download.pytorch.org/models/retinanet_resnet50_fpn_v2_coco-5905b1c5.pth" to /root/.cache/torch/hub/checkpoints/retinanet_resnet50_fpn_v2_coco-5905b1c5.pth
100% 146M/146M [00:00<00:00, 182MB/s]
RetinaNet(
(backbone): BackboneWithFPN(
(body): IntermediateLayerGetter(
(conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
(layer1): Sequential(
(0): Bottleneck(
(conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer2): Sequential(
(0): Bottleneck(
(conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer3): Sequential(
(0): Bottleneck(
(conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(4): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(5): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer4): Sequential(
(0): Bottleneck(
(conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
)
(fpn): FeaturePyramidNetwork(
(inner_blocks): ModuleList(
(0): Conv2dNormActivation(
(0): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1))
)
(1): Conv2dNormActivation(
(0): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1))
)
(2): Conv2dNormActivation(
(0): Conv2d(2048, 256, kernel_size=(1, 1), stride=(1, 1))
)
)
(layer_blocks): ModuleList(
(0-2): 3 x Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
)
)
(extra_blocks): LastLevelP6P7(
(p6): Conv2d(2048, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
(p7): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
)
)
)
(anchor_generator): AnchorGenerator()
(head): RetinaNetHead(
(classification_head): RetinaNetClassificationHead(
(conv): Sequential(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
)
(cls_logits): Conv2d(256, 45, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
)
(regression_head): RetinaNetRegressionHead(
(conv): Sequential(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
)
(bbox_reg): Conv2d(256, 36, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
)
)
(transform): GeneralizedRCNNTransform(
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
Resize(min_size=(800,), max_size=1333, mode='bilinear')
)
)
36,414,865 total parameters.
36,189,521 training parameters.
qt.qpa.xcb: could not connect to display
qt.qpa.plugin: Could not load the Qt platform plugin "xcb" in "/usr/local/lib/python3.12/dist-packages/cv2/qt/plugins" even though it was found.
This application failed to start because no Qt platform plugin could be initialized. Reinstalling the application may fix this problem.
Available platform plugins are: xcb.
Ahora pruebo a ejecutar app.py
In [17]:
!pip install gradio
Requirement already satisfied: gradio in /usr/local/lib/python3.12/dist-packages (5.18.0) Requirement already satisfied: aiofiles<24.0,>=22.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (23.2.1) Requirement already satisfied: anyio<5.0,>=3.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (4.10.0) Requirement already satisfied: fastapi<1.0,>=0.115.2 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.116.1) Requirement already satisfied: ffmpy in /usr/local/lib/python3.12/dist-packages (from gradio) (0.6.1) Requirement already satisfied: gradio-client==1.7.2 in /usr/local/lib/python3.12/dist-packages (from gradio) (1.7.2) Requirement already satisfied: httpx>=0.24.1 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.28.1) Requirement already satisfied: huggingface-hub>=0.28.1 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.34.4) Requirement already satisfied: jinja2<4.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (3.1.6) Requirement already satisfied: markupsafe~=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.1.5) Requirement already satisfied: numpy<3.0,>=1.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.0.2) Requirement already satisfied: orjson~=3.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (3.11.2) Requirement already satisfied: packaging in /usr/local/lib/python3.12/dist-packages (from gradio) (25.0) Requirement already satisfied: pandas<3.0,>=1.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.2.2) Requirement already satisfied: pillow<12.0,>=8.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (11.3.0) Requirement already satisfied: pydantic>=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.11.7) Requirement already satisfied: pydub in /usr/local/lib/python3.12/dist-packages (from gradio) (0.25.1) Requirement already satisfied: python-multipart>=0.0.18 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.0.20) Requirement already satisfied: pyyaml<7.0,>=5.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (6.0.2) Requirement already satisfied: ruff>=0.9.3 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.12.9) Requirement already satisfied: safehttpx<0.2.0,>=0.1.6 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.1.6) Requirement already satisfied: semantic-version~=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.10.0) Requirement already satisfied: starlette<1.0,>=0.40.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.47.2) Requirement already satisfied: tomlkit<0.14.0,>=0.12.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.13.3) Requirement already satisfied: typer<1.0,>=0.12 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.16.0) Requirement already satisfied: typing-extensions~=4.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (4.14.1) Requirement already satisfied: uvicorn>=0.14.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.35.0) Requirement already satisfied: fsspec in /usr/local/lib/python3.12/dist-packages (from gradio-client==1.7.2->gradio) (2025.3.0) Requirement already satisfied: websockets<16.0,>=10.0 in /usr/local/lib/python3.12/dist-packages (from gradio-client==1.7.2->gradio) (15.0.1) Requirement already satisfied: idna>=2.8 in /usr/local/lib/python3.12/dist-packages (from anyio<5.0,>=3.0->gradio) (3.10) Requirement already satisfied: sniffio>=1.1 in /usr/local/lib/python3.12/dist-packages (from anyio<5.0,>=3.0->gradio) (1.3.1) Requirement already satisfied: certifi in /usr/local/lib/python3.12/dist-packages (from httpx>=0.24.1->gradio) (2025.8.3) Requirement already satisfied: httpcore==1.* in /usr/local/lib/python3.12/dist-packages (from httpx>=0.24.1->gradio) (1.0.9) Requirement already satisfied: h11>=0.16 in /usr/local/lib/python3.12/dist-packages (from httpcore==1.*->httpx>=0.24.1->gradio) (0.16.0) Requirement already satisfied: filelock in /usr/local/lib/python3.12/dist-packages (from huggingface-hub>=0.28.1->gradio) (3.19.1) Requirement already satisfied: requests in /usr/local/lib/python3.12/dist-packages (from huggingface-hub>=0.28.1->gradio) (2.32.4) Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.12/dist-packages (from huggingface-hub>=0.28.1->gradio) (4.67.1) Requirement already satisfied: hf-xet<2.0.0,>=1.1.3 in /usr/local/lib/python3.12/dist-packages (from huggingface-hub>=0.28.1->gradio) (1.1.7) Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio) (2.9.0.post0) Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio) (2025.2) Requirement already satisfied: tzdata>=2022.7 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio) (2025.2) Requirement already satisfied: annotated-types>=0.6.0 in /usr/local/lib/python3.12/dist-packages (from pydantic>=2.0->gradio) (0.7.0) Requirement already satisfied: pydantic-core==2.33.2 in /usr/local/lib/python3.12/dist-packages (from pydantic>=2.0->gradio) (2.33.2) Requirement already satisfied: typing-inspection>=0.4.0 in /usr/local/lib/python3.12/dist-packages (from pydantic>=2.0->gradio) (0.4.1) Requirement already satisfied: click>=8.0.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio) (8.2.1) Requirement already satisfied: shellingham>=1.3.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio) (1.5.4) Requirement already satisfied: rich>=10.11.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio) (13.9.4) Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.12/dist-packages (from python-dateutil>=2.8.2->pandas<3.0,>=1.0->gradio) (1.17.0) Requirement already satisfied: markdown-it-py>=2.2.0 in /usr/local/lib/python3.12/dist-packages (from rich>=10.11.0->typer<1.0,>=0.12->gradio) (4.0.0) Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /usr/local/lib/python3.12/dist-packages (from rich>=10.11.0->typer<1.0,>=0.12->gradio) (2.19.2) Requirement already satisfied: charset_normalizer<4,>=2 in /usr/local/lib/python3.12/dist-packages (from requests->huggingface-hub>=0.28.1->gradio) (3.4.3) Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.12/dist-packages (from requests->huggingface-hub>=0.28.1->gradio) (2.5.0) Requirement already satisfied: mdurl~=0.1 in /usr/local/lib/python3.12/dist-packages (from markdown-it-py>=2.2.0->rich>=10.11.0->typer<1.0,>=0.12->gradio) (0.1.2)
In [18]:
!mkdir -p outputs
!cp best_model_79.pth outputs/
No sirve app.py, intento cambiar el código para que permita abrir la interfaz desde colab. También hay boleanos que hacen que se quede en un loop.
In [19]:
%%writefile /content/app.py
# app.py
import os
import cv2
import time
import torch
import gradio as gr
import numpy as np
# Make sure these are your local imports from your project.
from model import create_model
from config import NUM_CLASSES, DEVICE, CLASSES
# ----------------------------------------------------------------
# GLOBAL SETUP
# ----------------------------------------------------------------
# Create the model and load the best weights.
model = create_model(num_classes=NUM_CLASSES)
checkpoint = torch.load("outputs/best_model_79.pth", map_location=DEVICE)
model.load_state_dict(checkpoint["model_state_dict"])
model.to(DEVICE).eval()
# Create a colors array for each class index.
# (length matches len(CLASSES), including background if you wish).
COLORS = np.random.uniform(0, 255, size=(len(CLASSES), 3))
# COLORS = [
# (255, 255, 0), # Cyan - background
# (50, 0, 255), # Red - buffalo
# (147, 20, 255), # Pink - elephant
# (0, 255, 0), # Green - rhino
# (238, 130, 238), # Violet - zebra
# ]
# ----------------------------------------------------------------
# HELPER FUNCTIONS
# ----------------------------------------------------------------
def inference_on_image(orig_image: np.ndarray, resize_dim=None, threshold=0.25):
"""
Runs inference on a single image (OpenCV BGR or NumPy array).
- resize_dim: if not None, we resize to (resize_dim, resize_dim)
- threshold: detection confidence threshold
Returns: processed image with bounding boxes drawn.
"""
image = orig_image.copy()
# Optionally resize for inference.
if resize_dim is not None:
image = cv2.resize(image, (resize_dim, resize_dim))
# Convert BGR to RGB, normalize [0..1]
image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB).astype(np.float32) / 255.0
# Move channels to front (C,H,W)
image_tensor = torch.tensor(image_rgb.transpose(2, 0, 1), dtype=torch.float).unsqueeze(0).to(DEVICE)
start_time = time.time()
# Inference
with torch.no_grad():
outputs = model(image_tensor)
end_time = time.time()
# Get the current fps.
fps = 1 / (end_time - start_time)
fps_text = f"FPS: {fps:.2f}"
# Move outputs to CPU numpy
outputs = [{k: v.cpu() for k, v in t.items()} for t in outputs]
boxes = outputs[0]["boxes"].numpy()
scores = outputs[0]["scores"].numpy()
labels = outputs[0]["labels"].numpy().astype(int)
# Filter out boxes with low confidence
valid_idx = np.where(scores >= threshold)[0]
boxes = boxes[valid_idx].astype(int)
labels = labels[valid_idx]
# If we resized for inference, rescale boxes back to orig_image size
if resize_dim is not None:
h_orig, w_orig = orig_image.shape[:2]
h_new, w_new = resize_dim, resize_dim
# scale boxes
boxes[:, [0, 2]] = (boxes[:, [0, 2]] / w_new) * w_orig
boxes[:, [1, 3]] = (boxes[:, [1, 3]] / h_new) * h_orig
# Draw bounding boxes
for box, label_idx in zip(boxes, labels):
class_name = CLASSES[label_idx] if 0 <= label_idx < len(CLASSES) else str(label_idx)
color = COLORS[label_idx % len(COLORS)][::-1] # BGR color
cv2.rectangle(orig_image, (box[0], box[1]), (box[2], box[3]), color, 5)
cv2.putText(orig_image, class_name, (box[0], box[1] - 5), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0, 0, 0), 3)
cv2.putText(
orig_image,
fps_text,
(int((w_orig / 2) - 50), 30),
cv2.FONT_HERSHEY_SIMPLEX,
0.8,
(0, 255, 0),
2,
cv2.LINE_AA,
)
return orig_image, fps
def inference_on_frame(frame: np.ndarray, resize_dim=None, threshold=0.25):
"""
Same as inference_on_image but for a single video frame.
Returns the processed frame with bounding boxes.
"""
return inference_on_image(frame, resize_dim, threshold)
# ----------------------------------------------------------------
# GRADIO FUNCTIONS
# ----------------------------------------------------------------
def img_inf(image_path, resize_dim, threshold):
"""
Gradio function for image inference.
:param image_path: File path from Gradio (uploaded image).
:param model_name: Selected model from Radio (not used if only one model).
Returns: A NumPy image array with bounding boxes.
"""
if image_path is None:
return None # No image provided
orig_image = cv2.imread(image_path) # BGR
if orig_image is None:
return None # Error reading image
result_image, _ = inference_on_image(orig_image, resize_dim=resize_dim, threshold=threshold)
# Return the image in RGB for Gradio's display
result_image_rgb = cv2.cvtColor(result_image, cv2.COLOR_BGR2RGB)
return result_image_rgb
def vid_inf(video_path, resize_dim, threshold):
"""
Gradio function for video inference.
Processes each frame, draws bounding boxes, and writes to an output video.
Returns: (last_processed_frame, output_video_file_path)
"""
if video_path is None:
return None, None # No video provided
# Prepare input capture
cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
return None, None
# Create an output file path
os.makedirs("inference_outputs/videos", exist_ok=True)
out_video_path = os.path.join("inference_outputs/videos", "video_output.mp4")
# out_video_path = "video_output.mp4"
# Get video properties
width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
fps = cap.get(cv2.CAP_PROP_FPS)
fourcc = cv2.VideoWriter_fourcc(*"mp4v") # or 'XVID'
# If FPS is 0 (some weird container), default to something
if fps <= 0:
fps = 20.0
out_writer = cv2.VideoWriter(out_video_path, fourcc, fps, (width, height))
frame_count = 0
total_fps = 0
while True:
ret, frame = cap.read()
if not ret:
break
# Inference on frame
processed_frame, frame_fps = inference_on_frame(frame, resize_dim=resize_dim, threshold=threshold)
total_fps += frame_fps
frame_count += 1
# Write the processed frame
out_writer.write(processed_frame)
yield cv2.cvtColor(processed_frame, cv2.COLOR_BGR2RGB), None
avg_fps = total_fps / frame_count
cap.release()
out_writer.release()
print(f"Average FPS: {avg_fps:.3f}")
yield None, out_video_path
# ----------------------------------------------------------------
# BUILD THE GRADIO INTERFACES
# ----------------------------------------------------------------
# For demonstration, we define two possible model radio choices.
# You can ignore or expand this if you only use RetinaNet.
resize_dim = gr.Slider(100, 1024, value=640, label="Resize Dimension", info="Resize image to this dimension")
threshold = gr.Slider(0, 1, value=0.5, label="Threshold", info="Confidence threshold for detection")
inputs_image = gr.Image(type="filepath", label="Input Image")
outputs_image = gr.Image(type="numpy", label="Output Image")
interface_image = gr.Interface(
fn=img_inf,
inputs=[inputs_image, resize_dim, threshold],
outputs=outputs_image,
title="Image Inference",
description="Upload your photo, select a model, and see the results!",
examples=[["examples/buffalo.jpg"], ["examples/zebra.jpg"]],
cache_examples=False,
)
resize_dim = gr.Slider(100, 1024, value=640, label="Resize Dimension", info="Resize image to this dimension")
threshold = gr.Slider(0, 1, value=0.5, label="Threshold", info="Confidence threshold for detection")
input_video = gr.Video(label="Input Video")
# Output is a pair: (last_processed_frame, output_video_path)
output_frame = gr.Image(type="numpy", label="Output (Last Processed Frame)")
output_video_file = gr.Video(format="mp4", label="Output Video")
interface_video = gr.Interface(
fn=vid_inf,
inputs=[input_video, resize_dim, threshold],
outputs=[output_frame, output_video_file],
title="Video Inference",
description="Upload your video and see the processed output!",
examples=[["examples/elephants.mp4"], ["examples/rhino.mp4"]],
cache_examples=False,
)
# Combine them in a Tabbed Interface
# Lanzamiento de la app Gradio
# ----------------------------------------------------------------
demo = gr.TabbedInterface(
[interface_image, interface_video],
tab_names=["Image", "Video"],
title="FineTuning RetinaNet for Wildlife Animal Detection",
theme="gstaff/xkcd",
)
# Usar .queue() para procesar video de manera eficiente y lanzar en Colab
demo.queue().launch(share=True, inbrowser=True)
Overwriting /content/app.py
Aún hay errores con boleanos, se corrigió actualizando Gradio.
In [20]:
pip install --upgrade gradio gradio-client
Requirement already satisfied: gradio in /usr/local/lib/python3.12/dist-packages (5.18.0) Collecting gradio Downloading gradio-5.43.1-py3-none-any.whl.metadata (16 kB) Requirement already satisfied: gradio-client in /usr/local/lib/python3.12/dist-packages (1.7.2) Collecting gradio-client Downloading gradio_client-1.12.1-py3-none-any.whl.metadata (7.1 kB) Requirement already satisfied: aiofiles<25.0,>=22.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (23.2.1) Requirement already satisfied: anyio<5.0,>=3.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (4.10.0) Requirement already satisfied: brotli>=1.1.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (1.1.0) Requirement already satisfied: fastapi<1.0,>=0.115.2 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.116.1) Requirement already satisfied: ffmpy in /usr/local/lib/python3.12/dist-packages (from gradio) (0.6.1) Requirement already satisfied: groovy~=0.1 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.1.2) Requirement already satisfied: httpx<1.0,>=0.24.1 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.28.1) Requirement already satisfied: huggingface-hub<1.0,>=0.33.5 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.34.4) Requirement already satisfied: jinja2<4.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (3.1.6) Requirement already satisfied: markupsafe<4.0,>=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.1.5) Requirement already satisfied: numpy<3.0,>=1.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.0.2) Requirement already satisfied: orjson~=3.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (3.11.2) Requirement already satisfied: packaging in /usr/local/lib/python3.12/dist-packages (from gradio) (25.0) Requirement already satisfied: pandas<3.0,>=1.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.2.2) Requirement already satisfied: pillow<12.0,>=8.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (11.3.0) Requirement already satisfied: pydantic<2.12,>=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.11.7) Requirement already satisfied: pydub in /usr/local/lib/python3.12/dist-packages (from gradio) (0.25.1) Requirement already satisfied: python-multipart>=0.0.18 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.0.20) Requirement already satisfied: pyyaml<7.0,>=5.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (6.0.2) Requirement already satisfied: ruff>=0.9.3 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.12.9) Requirement already satisfied: safehttpx<0.2.0,>=0.1.6 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.1.6) Requirement already satisfied: semantic-version~=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.10.0) Requirement already satisfied: starlette<1.0,>=0.40.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.47.2) Requirement already satisfied: tomlkit<0.14.0,>=0.12.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.13.3) Requirement already satisfied: typer<1.0,>=0.12 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.16.0) Requirement already satisfied: typing-extensions~=4.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (4.14.1) Requirement already satisfied: uvicorn>=0.14.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.35.0) Requirement already satisfied: fsspec in /usr/local/lib/python3.12/dist-packages (from gradio-client) (2025.3.0) Requirement already satisfied: websockets<16.0,>=10.0 in /usr/local/lib/python3.12/dist-packages (from gradio-client) (15.0.1) Requirement already satisfied: idna>=2.8 in /usr/local/lib/python3.12/dist-packages (from anyio<5.0,>=3.0->gradio) (3.10) Requirement already satisfied: sniffio>=1.1 in /usr/local/lib/python3.12/dist-packages (from anyio<5.0,>=3.0->gradio) (1.3.1) Requirement already satisfied: certifi in /usr/local/lib/python3.12/dist-packages (from httpx<1.0,>=0.24.1->gradio) (2025.8.3) Requirement already satisfied: httpcore==1.* in /usr/local/lib/python3.12/dist-packages (from httpx<1.0,>=0.24.1->gradio) (1.0.9) Requirement already satisfied: h11>=0.16 in /usr/local/lib/python3.12/dist-packages (from httpcore==1.*->httpx<1.0,>=0.24.1->gradio) (0.16.0) Requirement already satisfied: filelock in /usr/local/lib/python3.12/dist-packages (from huggingface-hub<1.0,>=0.33.5->gradio) (3.19.1) Requirement already satisfied: requests in /usr/local/lib/python3.12/dist-packages (from huggingface-hub<1.0,>=0.33.5->gradio) (2.32.4) Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.12/dist-packages (from huggingface-hub<1.0,>=0.33.5->gradio) (4.67.1) Requirement already satisfied: hf-xet<2.0.0,>=1.1.3 in /usr/local/lib/python3.12/dist-packages (from huggingface-hub<1.0,>=0.33.5->gradio) (1.1.7) Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio) (2.9.0.post0) Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio) (2025.2) Requirement already satisfied: tzdata>=2022.7 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio) (2025.2) Requirement already satisfied: annotated-types>=0.6.0 in /usr/local/lib/python3.12/dist-packages (from pydantic<2.12,>=2.0->gradio) (0.7.0) Requirement already satisfied: pydantic-core==2.33.2 in /usr/local/lib/python3.12/dist-packages (from pydantic<2.12,>=2.0->gradio) (2.33.2) Requirement already satisfied: typing-inspection>=0.4.0 in /usr/local/lib/python3.12/dist-packages (from pydantic<2.12,>=2.0->gradio) (0.4.1) Requirement already satisfied: click>=8.0.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio) (8.2.1) Requirement already satisfied: shellingham>=1.3.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio) (1.5.4) Requirement already satisfied: rich>=10.11.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio) (13.9.4) Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.12/dist-packages (from python-dateutil>=2.8.2->pandas<3.0,>=1.0->gradio) (1.17.0) Requirement already satisfied: markdown-it-py>=2.2.0 in /usr/local/lib/python3.12/dist-packages (from rich>=10.11.0->typer<1.0,>=0.12->gradio) (4.0.0) Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /usr/local/lib/python3.12/dist-packages (from rich>=10.11.0->typer<1.0,>=0.12->gradio) (2.19.2) Requirement already satisfied: charset_normalizer<4,>=2 in /usr/local/lib/python3.12/dist-packages (from requests->huggingface-hub<1.0,>=0.33.5->gradio) (3.4.3) Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.12/dist-packages (from requests->huggingface-hub<1.0,>=0.33.5->gradio) (2.5.0) Requirement already satisfied: mdurl~=0.1 in /usr/local/lib/python3.12/dist-packages (from markdown-it-py>=2.2.0->rich>=10.11.0->typer<1.0,>=0.12->gradio) (0.1.2) Downloading gradio-5.43.1-py3-none-any.whl (59.6 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 59.6/59.6 MB 13.4 MB/s eta 0:00:00 Downloading gradio_client-1.12.1-py3-none-any.whl (324 kB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 324.6/324.6 kB 30.5 MB/s eta 0:00:00 Installing collected packages: gradio-client, gradio Attempting uninstall: gradio-client Found existing installation: gradio_client 1.7.2 Uninstalling gradio_client-1.7.2: Successfully uninstalled gradio_client-1.7.2 Attempting uninstall: gradio Found existing installation: gradio 5.18.0 Uninstalling gradio-5.18.0: Successfully uninstalled gradio-5.18.0 Successfully installed gradio-5.43.1 gradio-client-1.12.1
Finalmente se ejecuta app.py. Contiene una interfaz de reconocimiento de imágenes. Este se ejecuta con colab como servidor. Con esto se pueden identificar las especies de este conjunto de datos que se encuentren en cualquier imagen.
In [21]:
%cd /content/
!python app.py
/content
theme_schema%400.0.4.json: 12.6kB [00:00, 49.5MB/s]
* Running on local URL: http://127.0.0.1:7860
* Running on public URL: https://d20b1e69686cd0b3e0.gradio.live
This share link expires in 1 week. For free permanent hosting and GPU upgrades, run `gradio deploy` from the terminal in the working directory to deploy to Hugging Face Spaces (https://huggingface.co/spaces)
Keyboard interruption in main thread... closing server.
Traceback (most recent call last):
File "/usr/local/lib/python3.12/dist-packages/gradio/blocks.py", line 3158, in block_thread
time.sleep(0.1)
KeyboardInterrupt
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/content/app.py", line 237, in <module>
demo.queue().launch(share=True, inbrowser=True)
File "/usr/local/lib/python3.12/dist-packages/gradio/blocks.py", line 3055, in launch
self.block_thread()
File "/usr/local/lib/python3.12/dist-packages/gradio/blocks.py", line 3162, in block_thread
self.server.close()
File "/usr/local/lib/python3.12/dist-packages/gradio/http_server.py", line 69, in close
self.thread.join(timeout=5)
File "/usr/lib/python3.12/threading.py", line 1153, in join
self._wait_for_tstate_lock(timeout=max(timeout, 0))
File "/usr/lib/python3.12/threading.py", line 1169, in _wait_for_tstate_lock
if lock.acquire(block, timeout):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
KeyboardInterrupt
Killing tunnel 127.0.0.1:7860 <> https://d20b1e69686cd0b3e0.gradio.live
Ajusté el código para que las carpetas sólo se copien, por lo que aún tengo el modelo anterior disponible.
Ahora quiero comparar los modelos con un gráfico. Intento obtener el gráfico Precision-recall.
Ver el contenido de best_model_79.pth
In [24]:
import torch
# Cargar el archivo
checkpoint = torch.load("best_model_79.pth", map_location="cpu")
# Ver tipo y llaves
print(type(checkpoint))
print(checkpoint.keys() if isinstance(checkpoint, dict) else "No es diccionario")
<class 'dict'> dict_keys(['epoch', 'model_state_dict'])
Ver los pesos del modelo
In [25]:
state_dict = checkpoint if not isinstance(checkpoint, dict) else checkpoint['model_state_dict']
for k, v in state_dict.items():
print(k, v.shape)
backbone.body.conv1.weight torch.Size([64, 3, 7, 7]) backbone.body.bn1.weight torch.Size([64]) backbone.body.bn1.bias torch.Size([64]) backbone.body.bn1.running_mean torch.Size([64]) backbone.body.bn1.running_var torch.Size([64]) backbone.body.bn1.num_batches_tracked torch.Size([]) backbone.body.layer1.0.conv1.weight torch.Size([64, 64, 1, 1]) backbone.body.layer1.0.bn1.weight torch.Size([64]) backbone.body.layer1.0.bn1.bias torch.Size([64]) backbone.body.layer1.0.bn1.running_mean torch.Size([64]) backbone.body.layer1.0.bn1.running_var torch.Size([64]) backbone.body.layer1.0.bn1.num_batches_tracked torch.Size([]) backbone.body.layer1.0.conv2.weight torch.Size([64, 64, 3, 3]) backbone.body.layer1.0.bn2.weight torch.Size([64]) backbone.body.layer1.0.bn2.bias torch.Size([64]) backbone.body.layer1.0.bn2.running_mean torch.Size([64]) backbone.body.layer1.0.bn2.running_var torch.Size([64]) backbone.body.layer1.0.bn2.num_batches_tracked torch.Size([]) backbone.body.layer1.0.conv3.weight torch.Size([256, 64, 1, 1]) backbone.body.layer1.0.bn3.weight torch.Size([256]) backbone.body.layer1.0.bn3.bias torch.Size([256]) backbone.body.layer1.0.bn3.running_mean torch.Size([256]) backbone.body.layer1.0.bn3.running_var torch.Size([256]) backbone.body.layer1.0.bn3.num_batches_tracked torch.Size([]) backbone.body.layer1.0.downsample.0.weight torch.Size([256, 64, 1, 1]) backbone.body.layer1.0.downsample.1.weight torch.Size([256]) backbone.body.layer1.0.downsample.1.bias torch.Size([256]) backbone.body.layer1.0.downsample.1.running_mean torch.Size([256]) backbone.body.layer1.0.downsample.1.running_var torch.Size([256]) backbone.body.layer1.0.downsample.1.num_batches_tracked torch.Size([]) backbone.body.layer1.1.conv1.weight torch.Size([64, 256, 1, 1]) backbone.body.layer1.1.bn1.weight torch.Size([64]) backbone.body.layer1.1.bn1.bias torch.Size([64]) backbone.body.layer1.1.bn1.running_mean torch.Size([64]) backbone.body.layer1.1.bn1.running_var torch.Size([64]) backbone.body.layer1.1.bn1.num_batches_tracked torch.Size([]) backbone.body.layer1.1.conv2.weight torch.Size([64, 64, 3, 3]) backbone.body.layer1.1.bn2.weight torch.Size([64]) backbone.body.layer1.1.bn2.bias torch.Size([64]) backbone.body.layer1.1.bn2.running_mean torch.Size([64]) backbone.body.layer1.1.bn2.running_var torch.Size([64]) backbone.body.layer1.1.bn2.num_batches_tracked torch.Size([]) backbone.body.layer1.1.conv3.weight torch.Size([256, 64, 1, 1]) backbone.body.layer1.1.bn3.weight torch.Size([256]) backbone.body.layer1.1.bn3.bias torch.Size([256]) backbone.body.layer1.1.bn3.running_mean torch.Size([256]) backbone.body.layer1.1.bn3.running_var torch.Size([256]) backbone.body.layer1.1.bn3.num_batches_tracked torch.Size([]) backbone.body.layer1.2.conv1.weight torch.Size([64, 256, 1, 1]) backbone.body.layer1.2.bn1.weight torch.Size([64]) backbone.body.layer1.2.bn1.bias torch.Size([64]) backbone.body.layer1.2.bn1.running_mean torch.Size([64]) backbone.body.layer1.2.bn1.running_var torch.Size([64]) backbone.body.layer1.2.bn1.num_batches_tracked torch.Size([]) backbone.body.layer1.2.conv2.weight torch.Size([64, 64, 3, 3]) backbone.body.layer1.2.bn2.weight torch.Size([64]) backbone.body.layer1.2.bn2.bias torch.Size([64]) backbone.body.layer1.2.bn2.running_mean torch.Size([64]) backbone.body.layer1.2.bn2.running_var torch.Size([64]) backbone.body.layer1.2.bn2.num_batches_tracked torch.Size([]) backbone.body.layer1.2.conv3.weight torch.Size([256, 64, 1, 1]) backbone.body.layer1.2.bn3.weight torch.Size([256]) backbone.body.layer1.2.bn3.bias torch.Size([256]) backbone.body.layer1.2.bn3.running_mean torch.Size([256]) backbone.body.layer1.2.bn3.running_var torch.Size([256]) backbone.body.layer1.2.bn3.num_batches_tracked torch.Size([]) backbone.body.layer2.0.conv1.weight torch.Size([128, 256, 1, 1]) backbone.body.layer2.0.bn1.weight torch.Size([128]) backbone.body.layer2.0.bn1.bias torch.Size([128]) backbone.body.layer2.0.bn1.running_mean torch.Size([128]) backbone.body.layer2.0.bn1.running_var torch.Size([128]) backbone.body.layer2.0.bn1.num_batches_tracked torch.Size([]) backbone.body.layer2.0.conv2.weight torch.Size([128, 128, 3, 3]) backbone.body.layer2.0.bn2.weight torch.Size([128]) backbone.body.layer2.0.bn2.bias torch.Size([128]) backbone.body.layer2.0.bn2.running_mean torch.Size([128]) backbone.body.layer2.0.bn2.running_var torch.Size([128]) backbone.body.layer2.0.bn2.num_batches_tracked torch.Size([]) backbone.body.layer2.0.conv3.weight torch.Size([512, 128, 1, 1]) backbone.body.layer2.0.bn3.weight torch.Size([512]) backbone.body.layer2.0.bn3.bias torch.Size([512]) backbone.body.layer2.0.bn3.running_mean torch.Size([512]) backbone.body.layer2.0.bn3.running_var torch.Size([512]) backbone.body.layer2.0.bn3.num_batches_tracked torch.Size([]) backbone.body.layer2.0.downsample.0.weight torch.Size([512, 256, 1, 1]) backbone.body.layer2.0.downsample.1.weight torch.Size([512]) backbone.body.layer2.0.downsample.1.bias torch.Size([512]) backbone.body.layer2.0.downsample.1.running_mean torch.Size([512]) backbone.body.layer2.0.downsample.1.running_var torch.Size([512]) backbone.body.layer2.0.downsample.1.num_batches_tracked torch.Size([]) backbone.body.layer2.1.conv1.weight torch.Size([128, 512, 1, 1]) backbone.body.layer2.1.bn1.weight torch.Size([128]) backbone.body.layer2.1.bn1.bias torch.Size([128]) backbone.body.layer2.1.bn1.running_mean torch.Size([128]) backbone.body.layer2.1.bn1.running_var torch.Size([128]) backbone.body.layer2.1.bn1.num_batches_tracked torch.Size([]) backbone.body.layer2.1.conv2.weight torch.Size([128, 128, 3, 3]) backbone.body.layer2.1.bn2.weight torch.Size([128]) backbone.body.layer2.1.bn2.bias torch.Size([128]) backbone.body.layer2.1.bn2.running_mean torch.Size([128]) backbone.body.layer2.1.bn2.running_var torch.Size([128]) backbone.body.layer2.1.bn2.num_batches_tracked torch.Size([]) backbone.body.layer2.1.conv3.weight torch.Size([512, 128, 1, 1]) backbone.body.layer2.1.bn3.weight torch.Size([512]) backbone.body.layer2.1.bn3.bias torch.Size([512]) backbone.body.layer2.1.bn3.running_mean torch.Size([512]) backbone.body.layer2.1.bn3.running_var torch.Size([512]) backbone.body.layer2.1.bn3.num_batches_tracked torch.Size([]) backbone.body.layer2.2.conv1.weight torch.Size([128, 512, 1, 1]) backbone.body.layer2.2.bn1.weight torch.Size([128]) backbone.body.layer2.2.bn1.bias torch.Size([128]) backbone.body.layer2.2.bn1.running_mean torch.Size([128]) backbone.body.layer2.2.bn1.running_var torch.Size([128]) backbone.body.layer2.2.bn1.num_batches_tracked torch.Size([]) backbone.body.layer2.2.conv2.weight torch.Size([128, 128, 3, 3]) backbone.body.layer2.2.bn2.weight torch.Size([128]) backbone.body.layer2.2.bn2.bias torch.Size([128]) backbone.body.layer2.2.bn2.running_mean torch.Size([128]) backbone.body.layer2.2.bn2.running_var torch.Size([128]) backbone.body.layer2.2.bn2.num_batches_tracked torch.Size([]) backbone.body.layer2.2.conv3.weight torch.Size([512, 128, 1, 1]) backbone.body.layer2.2.bn3.weight torch.Size([512]) backbone.body.layer2.2.bn3.bias torch.Size([512]) backbone.body.layer2.2.bn3.running_mean torch.Size([512]) backbone.body.layer2.2.bn3.running_var torch.Size([512]) backbone.body.layer2.2.bn3.num_batches_tracked torch.Size([]) backbone.body.layer2.3.conv1.weight torch.Size([128, 512, 1, 1]) backbone.body.layer2.3.bn1.weight torch.Size([128]) backbone.body.layer2.3.bn1.bias torch.Size([128]) backbone.body.layer2.3.bn1.running_mean torch.Size([128]) backbone.body.layer2.3.bn1.running_var torch.Size([128]) backbone.body.layer2.3.bn1.num_batches_tracked torch.Size([]) backbone.body.layer2.3.conv2.weight torch.Size([128, 128, 3, 3]) backbone.body.layer2.3.bn2.weight torch.Size([128]) backbone.body.layer2.3.bn2.bias torch.Size([128]) backbone.body.layer2.3.bn2.running_mean torch.Size([128]) backbone.body.layer2.3.bn2.running_var torch.Size([128]) backbone.body.layer2.3.bn2.num_batches_tracked torch.Size([]) backbone.body.layer2.3.conv3.weight torch.Size([512, 128, 1, 1]) backbone.body.layer2.3.bn3.weight torch.Size([512]) backbone.body.layer2.3.bn3.bias torch.Size([512]) backbone.body.layer2.3.bn3.running_mean torch.Size([512]) backbone.body.layer2.3.bn3.running_var torch.Size([512]) backbone.body.layer2.3.bn3.num_batches_tracked torch.Size([]) backbone.body.layer3.0.conv1.weight torch.Size([256, 512, 1, 1]) backbone.body.layer3.0.bn1.weight torch.Size([256]) backbone.body.layer3.0.bn1.bias torch.Size([256]) backbone.body.layer3.0.bn1.running_mean torch.Size([256]) backbone.body.layer3.0.bn1.running_var torch.Size([256]) backbone.body.layer3.0.bn1.num_batches_tracked torch.Size([]) backbone.body.layer3.0.conv2.weight torch.Size([256, 256, 3, 3]) backbone.body.layer3.0.bn2.weight torch.Size([256]) backbone.body.layer3.0.bn2.bias torch.Size([256]) backbone.body.layer3.0.bn2.running_mean torch.Size([256]) backbone.body.layer3.0.bn2.running_var torch.Size([256]) backbone.body.layer3.0.bn2.num_batches_tracked torch.Size([]) backbone.body.layer3.0.conv3.weight torch.Size([1024, 256, 1, 1]) backbone.body.layer3.0.bn3.weight torch.Size([1024]) backbone.body.layer3.0.bn3.bias torch.Size([1024]) backbone.body.layer3.0.bn3.running_mean torch.Size([1024]) backbone.body.layer3.0.bn3.running_var torch.Size([1024]) backbone.body.layer3.0.bn3.num_batches_tracked torch.Size([]) backbone.body.layer3.0.downsample.0.weight torch.Size([1024, 512, 1, 1]) backbone.body.layer3.0.downsample.1.weight torch.Size([1024]) backbone.body.layer3.0.downsample.1.bias torch.Size([1024]) backbone.body.layer3.0.downsample.1.running_mean torch.Size([1024]) backbone.body.layer3.0.downsample.1.running_var torch.Size([1024]) backbone.body.layer3.0.downsample.1.num_batches_tracked torch.Size([]) backbone.body.layer3.1.conv1.weight torch.Size([256, 1024, 1, 1]) backbone.body.layer3.1.bn1.weight torch.Size([256]) backbone.body.layer3.1.bn1.bias torch.Size([256]) backbone.body.layer3.1.bn1.running_mean torch.Size([256]) backbone.body.layer3.1.bn1.running_var torch.Size([256]) backbone.body.layer3.1.bn1.num_batches_tracked torch.Size([]) backbone.body.layer3.1.conv2.weight torch.Size([256, 256, 3, 3]) backbone.body.layer3.1.bn2.weight torch.Size([256]) backbone.body.layer3.1.bn2.bias torch.Size([256]) backbone.body.layer3.1.bn2.running_mean torch.Size([256]) backbone.body.layer3.1.bn2.running_var torch.Size([256]) backbone.body.layer3.1.bn2.num_batches_tracked torch.Size([]) backbone.body.layer3.1.conv3.weight torch.Size([1024, 256, 1, 1]) backbone.body.layer3.1.bn3.weight torch.Size([1024]) backbone.body.layer3.1.bn3.bias torch.Size([1024]) backbone.body.layer3.1.bn3.running_mean torch.Size([1024]) backbone.body.layer3.1.bn3.running_var torch.Size([1024]) backbone.body.layer3.1.bn3.num_batches_tracked torch.Size([]) backbone.body.layer3.2.conv1.weight torch.Size([256, 1024, 1, 1]) backbone.body.layer3.2.bn1.weight torch.Size([256]) backbone.body.layer3.2.bn1.bias torch.Size([256]) backbone.body.layer3.2.bn1.running_mean torch.Size([256]) backbone.body.layer3.2.bn1.running_var torch.Size([256]) backbone.body.layer3.2.bn1.num_batches_tracked torch.Size([]) backbone.body.layer3.2.conv2.weight torch.Size([256, 256, 3, 3]) backbone.body.layer3.2.bn2.weight torch.Size([256]) backbone.body.layer3.2.bn2.bias torch.Size([256]) backbone.body.layer3.2.bn2.running_mean torch.Size([256]) backbone.body.layer3.2.bn2.running_var torch.Size([256]) backbone.body.layer3.2.bn2.num_batches_tracked torch.Size([]) backbone.body.layer3.2.conv3.weight torch.Size([1024, 256, 1, 1]) backbone.body.layer3.2.bn3.weight torch.Size([1024]) backbone.body.layer3.2.bn3.bias torch.Size([1024]) backbone.body.layer3.2.bn3.running_mean torch.Size([1024]) backbone.body.layer3.2.bn3.running_var torch.Size([1024]) backbone.body.layer3.2.bn3.num_batches_tracked torch.Size([]) backbone.body.layer3.3.conv1.weight torch.Size([256, 1024, 1, 1]) backbone.body.layer3.3.bn1.weight torch.Size([256]) backbone.body.layer3.3.bn1.bias torch.Size([256]) backbone.body.layer3.3.bn1.running_mean torch.Size([256]) backbone.body.layer3.3.bn1.running_var torch.Size([256]) backbone.body.layer3.3.bn1.num_batches_tracked torch.Size([]) backbone.body.layer3.3.conv2.weight torch.Size([256, 256, 3, 3]) backbone.body.layer3.3.bn2.weight torch.Size([256]) backbone.body.layer3.3.bn2.bias torch.Size([256]) backbone.body.layer3.3.bn2.running_mean torch.Size([256]) backbone.body.layer3.3.bn2.running_var torch.Size([256]) backbone.body.layer3.3.bn2.num_batches_tracked torch.Size([]) backbone.body.layer3.3.conv3.weight torch.Size([1024, 256, 1, 1]) backbone.body.layer3.3.bn3.weight torch.Size([1024]) backbone.body.layer3.3.bn3.bias torch.Size([1024]) backbone.body.layer3.3.bn3.running_mean torch.Size([1024]) backbone.body.layer3.3.bn3.running_var torch.Size([1024]) backbone.body.layer3.3.bn3.num_batches_tracked torch.Size([]) backbone.body.layer3.4.conv1.weight torch.Size([256, 1024, 1, 1]) backbone.body.layer3.4.bn1.weight torch.Size([256]) backbone.body.layer3.4.bn1.bias torch.Size([256]) backbone.body.layer3.4.bn1.running_mean torch.Size([256]) backbone.body.layer3.4.bn1.running_var torch.Size([256]) backbone.body.layer3.4.bn1.num_batches_tracked torch.Size([]) backbone.body.layer3.4.conv2.weight torch.Size([256, 256, 3, 3]) backbone.body.layer3.4.bn2.weight torch.Size([256]) backbone.body.layer3.4.bn2.bias torch.Size([256]) backbone.body.layer3.4.bn2.running_mean torch.Size([256]) backbone.body.layer3.4.bn2.running_var torch.Size([256]) backbone.body.layer3.4.bn2.num_batches_tracked torch.Size([]) backbone.body.layer3.4.conv3.weight torch.Size([1024, 256, 1, 1]) backbone.body.layer3.4.bn3.weight torch.Size([1024]) backbone.body.layer3.4.bn3.bias torch.Size([1024]) backbone.body.layer3.4.bn3.running_mean torch.Size([1024]) backbone.body.layer3.4.bn3.running_var torch.Size([1024]) backbone.body.layer3.4.bn3.num_batches_tracked torch.Size([]) backbone.body.layer3.5.conv1.weight torch.Size([256, 1024, 1, 1]) backbone.body.layer3.5.bn1.weight torch.Size([256]) backbone.body.layer3.5.bn1.bias torch.Size([256]) backbone.body.layer3.5.bn1.running_mean torch.Size([256]) backbone.body.layer3.5.bn1.running_var torch.Size([256]) backbone.body.layer3.5.bn1.num_batches_tracked torch.Size([]) backbone.body.layer3.5.conv2.weight torch.Size([256, 256, 3, 3]) backbone.body.layer3.5.bn2.weight torch.Size([256]) backbone.body.layer3.5.bn2.bias torch.Size([256]) backbone.body.layer3.5.bn2.running_mean torch.Size([256]) backbone.body.layer3.5.bn2.running_var torch.Size([256]) backbone.body.layer3.5.bn2.num_batches_tracked torch.Size([]) backbone.body.layer3.5.conv3.weight torch.Size([1024, 256, 1, 1]) backbone.body.layer3.5.bn3.weight torch.Size([1024]) backbone.body.layer3.5.bn3.bias torch.Size([1024]) backbone.body.layer3.5.bn3.running_mean torch.Size([1024]) backbone.body.layer3.5.bn3.running_var torch.Size([1024]) backbone.body.layer3.5.bn3.num_batches_tracked torch.Size([]) backbone.body.layer4.0.conv1.weight torch.Size([512, 1024, 1, 1]) backbone.body.layer4.0.bn1.weight torch.Size([512]) backbone.body.layer4.0.bn1.bias torch.Size([512]) backbone.body.layer4.0.bn1.running_mean torch.Size([512]) backbone.body.layer4.0.bn1.running_var torch.Size([512]) backbone.body.layer4.0.bn1.num_batches_tracked torch.Size([]) backbone.body.layer4.0.conv2.weight torch.Size([512, 512, 3, 3]) backbone.body.layer4.0.bn2.weight torch.Size([512]) backbone.body.layer4.0.bn2.bias torch.Size([512]) backbone.body.layer4.0.bn2.running_mean torch.Size([512]) backbone.body.layer4.0.bn2.running_var torch.Size([512]) backbone.body.layer4.0.bn2.num_batches_tracked torch.Size([]) backbone.body.layer4.0.conv3.weight torch.Size([2048, 512, 1, 1]) backbone.body.layer4.0.bn3.weight torch.Size([2048]) backbone.body.layer4.0.bn3.bias torch.Size([2048]) backbone.body.layer4.0.bn3.running_mean torch.Size([2048]) backbone.body.layer4.0.bn3.running_var torch.Size([2048]) backbone.body.layer4.0.bn3.num_batches_tracked torch.Size([]) backbone.body.layer4.0.downsample.0.weight torch.Size([2048, 1024, 1, 1]) backbone.body.layer4.0.downsample.1.weight torch.Size([2048]) backbone.body.layer4.0.downsample.1.bias torch.Size([2048]) backbone.body.layer4.0.downsample.1.running_mean torch.Size([2048]) backbone.body.layer4.0.downsample.1.running_var torch.Size([2048]) backbone.body.layer4.0.downsample.1.num_batches_tracked torch.Size([]) backbone.body.layer4.1.conv1.weight torch.Size([512, 2048, 1, 1]) backbone.body.layer4.1.bn1.weight torch.Size([512]) backbone.body.layer4.1.bn1.bias torch.Size([512]) backbone.body.layer4.1.bn1.running_mean torch.Size([512]) backbone.body.layer4.1.bn1.running_var torch.Size([512]) backbone.body.layer4.1.bn1.num_batches_tracked torch.Size([]) backbone.body.layer4.1.conv2.weight torch.Size([512, 512, 3, 3]) backbone.body.layer4.1.bn2.weight torch.Size([512]) backbone.body.layer4.1.bn2.bias torch.Size([512]) backbone.body.layer4.1.bn2.running_mean torch.Size([512]) backbone.body.layer4.1.bn2.running_var torch.Size([512]) backbone.body.layer4.1.bn2.num_batches_tracked torch.Size([]) backbone.body.layer4.1.conv3.weight torch.Size([2048, 512, 1, 1]) backbone.body.layer4.1.bn3.weight torch.Size([2048]) backbone.body.layer4.1.bn3.bias torch.Size([2048]) backbone.body.layer4.1.bn3.running_mean torch.Size([2048]) backbone.body.layer4.1.bn3.running_var torch.Size([2048]) backbone.body.layer4.1.bn3.num_batches_tracked torch.Size([]) backbone.body.layer4.2.conv1.weight torch.Size([512, 2048, 1, 1]) backbone.body.layer4.2.bn1.weight torch.Size([512]) backbone.body.layer4.2.bn1.bias torch.Size([512]) backbone.body.layer4.2.bn1.running_mean torch.Size([512]) backbone.body.layer4.2.bn1.running_var torch.Size([512]) backbone.body.layer4.2.bn1.num_batches_tracked torch.Size([]) backbone.body.layer4.2.conv2.weight torch.Size([512, 512, 3, 3]) backbone.body.layer4.2.bn2.weight torch.Size([512]) backbone.body.layer4.2.bn2.bias torch.Size([512]) backbone.body.layer4.2.bn2.running_mean torch.Size([512]) backbone.body.layer4.2.bn2.running_var torch.Size([512]) backbone.body.layer4.2.bn2.num_batches_tracked torch.Size([]) backbone.body.layer4.2.conv3.weight torch.Size([2048, 512, 1, 1]) backbone.body.layer4.2.bn3.weight torch.Size([2048]) backbone.body.layer4.2.bn3.bias torch.Size([2048]) backbone.body.layer4.2.bn3.running_mean torch.Size([2048]) backbone.body.layer4.2.bn3.running_var torch.Size([2048]) backbone.body.layer4.2.bn3.num_batches_tracked torch.Size([]) backbone.fpn.inner_blocks.0.0.weight torch.Size([256, 512, 1, 1]) backbone.fpn.inner_blocks.0.0.bias torch.Size([256]) backbone.fpn.inner_blocks.1.0.weight torch.Size([256, 1024, 1, 1]) backbone.fpn.inner_blocks.1.0.bias torch.Size([256]) backbone.fpn.inner_blocks.2.0.weight torch.Size([256, 2048, 1, 1]) backbone.fpn.inner_blocks.2.0.bias torch.Size([256]) backbone.fpn.layer_blocks.0.0.weight torch.Size([256, 256, 3, 3]) backbone.fpn.layer_blocks.0.0.bias torch.Size([256]) backbone.fpn.layer_blocks.1.0.weight torch.Size([256, 256, 3, 3]) backbone.fpn.layer_blocks.1.0.bias torch.Size([256]) backbone.fpn.layer_blocks.2.0.weight torch.Size([256, 256, 3, 3]) backbone.fpn.layer_blocks.2.0.bias torch.Size([256]) backbone.fpn.extra_blocks.p6.weight torch.Size([256, 2048, 3, 3]) backbone.fpn.extra_blocks.p6.bias torch.Size([256]) backbone.fpn.extra_blocks.p7.weight torch.Size([256, 256, 3, 3]) backbone.fpn.extra_blocks.p7.bias torch.Size([256]) head.classification_head.conv.0.0.weight torch.Size([256, 256, 3, 3]) head.classification_head.conv.0.1.weight torch.Size([256]) head.classification_head.conv.0.1.bias torch.Size([256]) head.classification_head.conv.1.0.weight torch.Size([256, 256, 3, 3]) head.classification_head.conv.1.1.weight torch.Size([256]) head.classification_head.conv.1.1.bias torch.Size([256]) head.classification_head.conv.2.0.weight torch.Size([256, 256, 3, 3]) head.classification_head.conv.2.1.weight torch.Size([256]) head.classification_head.conv.2.1.bias torch.Size([256]) head.classification_head.conv.3.0.weight torch.Size([256, 256, 3, 3]) head.classification_head.conv.3.1.weight torch.Size([256]) head.classification_head.conv.3.1.bias torch.Size([256]) head.classification_head.cls_logits.weight torch.Size([45, 256, 3, 3]) head.classification_head.cls_logits.bias torch.Size([45]) head.regression_head.conv.0.0.weight torch.Size([256, 256, 3, 3]) head.regression_head.conv.0.1.weight torch.Size([256]) head.regression_head.conv.0.1.bias torch.Size([256]) head.regression_head.conv.1.0.weight torch.Size([256, 256, 3, 3]) head.regression_head.conv.1.1.weight torch.Size([256]) head.regression_head.conv.1.1.bias torch.Size([256]) head.regression_head.conv.2.0.weight torch.Size([256, 256, 3, 3]) head.regression_head.conv.2.1.weight torch.Size([256]) head.regression_head.conv.2.1.bias torch.Size([256]) head.regression_head.conv.3.0.weight torch.Size([256, 256, 3, 3]) head.regression_head.conv.3.1.weight torch.Size([256]) head.regression_head.conv.3.1.bias torch.Size([256]) head.regression_head.bbox_reg.weight torch.Size([36, 256, 3, 3]) head.regression_head.bbox_reg.bias torch.Size([36])
Cargar el modelo para usarlo
In [26]:
from model import create_model #función de creación de modelo
from config import NUM_CLASSES, DEVICE
model = create_model(num_classes=NUM_CLASSES)
checkpoint = torch.load("best_model_79.pth", map_location=DEVICE)
model.load_state_dict(checkpoint['model_state_dict'])
model.to(DEVICE).eval()
Out[26]:
RetinaNet(
(backbone): BackboneWithFPN(
(body): IntermediateLayerGetter(
(conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
(layer1): Sequential(
(0): Bottleneck(
(conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer2): Sequential(
(0): Bottleneck(
(conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer3): Sequential(
(0): Bottleneck(
(conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(4): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(5): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer4): Sequential(
(0): Bottleneck(
(conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
)
(fpn): FeaturePyramidNetwork(
(inner_blocks): ModuleList(
(0): Conv2dNormActivation(
(0): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1))
)
(1): Conv2dNormActivation(
(0): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1))
)
(2): Conv2dNormActivation(
(0): Conv2d(2048, 256, kernel_size=(1, 1), stride=(1, 1))
)
)
(layer_blocks): ModuleList(
(0-2): 3 x Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
)
)
(extra_blocks): LastLevelP6P7(
(p6): Conv2d(2048, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
(p7): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
)
)
)
(anchor_generator): AnchorGenerator()
(head): RetinaNetHead(
(classification_head): RetinaNetClassificationHead(
(conv): Sequential(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
)
(cls_logits): Conv2d(256, 45, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
)
(regression_head): RetinaNetRegressionHead(
(conv): Sequential(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): GroupNorm(32, 256, eps=1e-05, affine=True)
(2): ReLU(inplace=True)
)
)
(bbox_reg): Conv2d(256, 36, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
)
)
(transform): GeneralizedRCNNTransform(
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
Resize(min_size=(800,), max_size=1333, mode='bilinear')
)
)
usar customdatasetd
In [27]:
from torch.utils.data import DataLoader
from datasets import create_valid_dataset #
from config import BATCH_SIZE, DEVICE
valid_dataset = create_valid_dataset("/content/datasets/african-wildlife/val")
valid_loader = DataLoader(valid_dataset, batch_size=1, shuffle=False)
/usr/local/lib/python3.12/dist-packages/albumentations/core/composition.py:331: UserWarning: Got processor for bboxes, but no transform to process it. self._set_keys()
Tras error, acceder a los labels de targets.
In [30]:
print(targets)
{'boxes': tensor([[[358.4000, 424.3309, 395.2000, 425.3309]]]), 'labels': tensor([[1]]), 'area': tensor([[36.8000]]), 'iscrowd': tensor([[0]]), 'image_id': tensor([[0]])}
In [31]:
true_labels = targets['labels'].squeeze(0).cpu().numpy()
In [32]:
true_boxes = targets['boxes'].squeeze(0).cpu().numpy()
Eliminar error Keyerror
In [33]:
all_scores = []
all_labels = []
for images, targets in valid_loader:
images = images.to(DEVICE)
with torch.no_grad():
outputs = model(images)
pred_scores = outputs[0]['scores'].cpu().numpy()
pred_labels = outputs[0]['labels'].cpu().numpy()
true_labels = targets['labels'].squeeze(0).cpu().numpy()
all_scores.extend(pred_scores)
all_labels.extend(true_labels)
Separar clases (especies)
In [34]:
from sklearn.metrics import precision_recall_curve, average_precision_score
import matplotlib.pyplot as plt
import numpy as np
# Lista de clases
classes = ["buffalo", "elephant", "rhino", "zebra"]
n_classes = len(classes)
# Binarizar etiquetas reales
y_true_bin = label_binarize(all_labels, classes=range(1, n_classes+1)) # ajusta según tu index
Calcular Precision-Recall y AP por clase
In [35]:
plt.figure(figsize=(8,6))
for i, class_name in enumerate(classes):
# Scores para esta clase
y_scores_class = np.array([s if l == i+1 else 0 for s,l in zip(all_scores, all_labels)])
y_true_class = y_true_bin[:, i]
precision, recall, _ = precision_recall_curve(y_true_class, y_scores_class)
ap = average_precision_score(y_true_class, y_scores_class)
plt.plot(recall, precision, label=f"{class_name} (AP={ap:.2f})")
plt.xlabel("Recall")
plt.ylabel("Precision")
plt.title("Precision-Recall curve por clase")
plt.legend()
plt.show()
El gráfico no salió bien, asume que el modelo es perfecto. Hice varias pruebas pero al final me rendí X_X
A partir de este punto sólo quedan intentos fallidos. No los remuevo para seguir intentando corregirlos algún día.
In [38]:
for images, targets in valid_loader:
images = list(img.to(DEVICE) for img in images)
with torch.no_grad():
outputs = model(images)
# Aquí targets es lista de diccionarios
for output, target in zip(outputs, targets):
# output es diccionario de predicciones
scores = output['scores'].cpu().numpy()
labels_pred = output['labels'].cpu().numpy()
# target es diccionario de ground-truth
labels_true = target['labels'].cpu().numpy().flatten() # <- flatten porque puede ser [[1]]
all_scores.extend(scores)
all_labels.extend(labels_true)
--------------------------------------------------------------------------- TypeError Traceback (most recent call last) /tmp/ipython-input-2908744283.py in <cell line: 0>() 12 13 # target es diccionario de ground-truth ---> 14 labels_true = target['labels'].cpu().numpy().flatten() # <- flatten porque puede ser [[1]] 15 16 all_scores.extend(scores) TypeError: string indices must be integers, not 'str'
No me sale. Intento la matriz de confusión (tampoco salió)
Exportación a Onnx.
In [ ]:
import torch
import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay
In [ ]:
!pip install onnx
Collecting onnx Downloading onnx-1.18.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (6.9 kB) Requirement already satisfied: numpy>=1.22 in /usr/local/lib/python3.12/dist-packages (from onnx) (2.0.2) Requirement already satisfied: protobuf>=4.25.1 in /usr/local/lib/python3.12/dist-packages (from onnx) (5.29.5) Requirement already satisfied: typing_extensions>=4.7.1 in /usr/local/lib/python3.12/dist-packages (from onnx) (4.14.1) Downloading onnx-1.18.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (17.6 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 17.6/17.6 MB 110.7 MB/s eta 0:00:00 Installing collected packages: onnx Successfully installed onnx-1.18.0
In [ ]:
!pip install onnxruntime
Collecting onnxruntime Downloading onnxruntime-1.22.1-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (4.9 kB) Collecting coloredlogs (from onnxruntime) Downloading coloredlogs-15.0.1-py2.py3-none-any.whl.metadata (12 kB) Requirement already satisfied: flatbuffers in /usr/local/lib/python3.12/dist-packages (from onnxruntime) (25.2.10) Requirement already satisfied: numpy>=1.21.6 in /usr/local/lib/python3.12/dist-packages (from onnxruntime) (2.0.2) Requirement already satisfied: packaging in /usr/local/lib/python3.12/dist-packages (from onnxruntime) (25.0) Requirement already satisfied: protobuf in /usr/local/lib/python3.12/dist-packages (from onnxruntime) (5.29.5) Requirement already satisfied: sympy in /usr/local/lib/python3.12/dist-packages (from onnxruntime) (1.13.1) Collecting humanfriendly>=9.1 (from coloredlogs->onnxruntime) Downloading humanfriendly-10.0-py2.py3-none-any.whl.metadata (9.2 kB) Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.12/dist-packages (from sympy->onnxruntime) (1.3.0) Downloading onnxruntime-1.22.1-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (16.5 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 16.5/16.5 MB 109.8 MB/s eta 0:00:00 Downloading coloredlogs-15.0.1-py2.py3-none-any.whl (46 kB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 46.0/46.0 kB 4.6 MB/s eta 0:00:00 Downloading humanfriendly-10.0-py2.py3-none-any.whl (86 kB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 86.8/86.8 kB 9.8 MB/s eta 0:00:00 Installing collected packages: humanfriendly, coloredlogs, onnxruntime Successfully installed coloredlogs-15.0.1 humanfriendly-10.0 onnxruntime-1.22.1
In [ ]:
!python export.py
/usr/local/lib/python3.12/dist-packages/torch/nn/functional.py:4624: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). * torch.tensor(scale_factors[i], dtype=torch.float32) /usr/local/lib/python3.12/dist-packages/torchvision/ops/boxes.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). boxes_x = torch.min(boxes_x, torch.tensor(width, dtype=boxes.dtype, device=boxes.device)) /usr/local/lib/python3.12/dist-packages/torchvision/ops/boxes.py:168: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). boxes_y = torch.min(boxes_y, torch.tensor(height, dtype=boxes.dtype, device=boxes.device)) /usr/local/lib/python3.12/dist-packages/torchvision/models/detection/transform.py:308: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). torch.tensor(s, dtype=torch.float32, device=boxes.device) /usr/local/lib/python3.12/dist-packages/torchvision/models/detection/transform.py:309: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). / torch.tensor(s_orig, dtype=torch.float32, device=boxes.device) Model exported to outputs/retinanet.onnx
In [ ]:
import onnxruntime as ort
import torch
import numpy as np
import cv2
# Ruta del modelo ONNX
onnx_model_path = "outputs/retinanet.onnx"
# Crear sesión ONNX
ort_session = ort.InferenceSession(onnx_model_path)
# Prepara una imagen de prueba
image_path = "/content/datasets/african-wildlife/train/images/1 (106).jpg"
img = cv2.imread(image_path)
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img = cv2.resize(img, (640, 640)) # tamaño usado en export.py
img = img.astype(np.float32) / 255.0
# Cambiar HWC -> CHW
img = np.transpose(img, (2, 0, 1))
img = np.expand_dims(img, axis=0) # añadir batch dimension: 1x3x640x640
# Convertir a float32
img = img.astype(np.float32)
# Ejecutar inferencia
outputs = ort_session.run(None, {"images": img})
# outputs es una lista: [boxes, scores, labels]
boxes, scores, labels = outputs
print("Boxes:", boxes.shape)
print("Scores:", scores.shape)
print("Labels:", labels.shape)
# Ejemplo: mostrar cajas con OpenCV
for i in range(len(boxes)):
if scores[i] > 0.5: # umbral de confianza
x1, y1, x2, y2 = boxes[i].astype(int)
label = int(labels[i])
cv2.rectangle(img[0].transpose(1,2,0) * 255, (x1, y1), (x2, y2), (0, 0, 255), 2)
Boxes: (4, 4) Scores: (4,) Labels: (4,)
In [ ]:
import onnxruntime as ort
import cv2
import numpy as np
# -----------------------------
# Config
# -----------------------------
onnx_model_path = "outputs/retinanet.onnx"
image_path = "/content/datasets/african-wildlife/train/images/1 (106).jpg"
input_size = 640 # mismo tamaño que usaste al exportar
CLASSES = ["buffalo", "zebra", "elephant", "rhino"] # reemplaza con tus clases reales
# -----------------------------
# 1️⃣ Cargar la imagen
# -----------------------------
img = cv2.imread(image_path)
img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img_resized = cv2.resize(img_rgb, (input_size, input_size))
img_normalized = img_resized.astype(np.float32) / 255.0
# Convertir a (1, 3, H, W)
input_tensor = np.transpose(img_normalized, (2, 0, 1))[np.newaxis, :, :, :]
# -----------------------------
# 2️⃣ Cargar el modelo ONNX
# -----------------------------
ort_session = ort.InferenceSession(onnx_model_path)
# -----------------------------
# 3️⃣ Realizar inferencia
# -----------------------------
outputs = ort_session.run(None, {"images": input_tensor})
boxes = outputs[0] # shape: [num_detections, 4]
scores = outputs[1] # shape: [num_detections]
labels = outputs[2] # shape: [num_detections]
# -----------------------------
# 4️⃣ Dibujar resultados
# -----------------------------
for box, score, label in zip(boxes, scores, labels):
if score < 0.3: # umbral de confianza
continue
x1, y1, x2, y2 = box
x1 = int(x1 * img.shape[1] / input_size)
x2 = int(x2 * img.shape[1] / input_size)
y1 = int(y1 * img.shape[0] / input_size)
y2 = int(y2 * img.shape[0] / input_size)
class_name = CLASSES[int(label) - 1] # si tu clase empieza en 1
cv2.rectangle(img, (x1, y1), (x2, y2), (0, 255, 0), 2)
cv2.putText(img, f"{class_name}: {score:.2f}", (x1, max(y1-10, 0)),
cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
# Mostrar imagen
cv2.imshow("Detections", img)
cv2.waitKey(0)
cv2.destroyAllWindows()
--------------------------------------------------------------------------- DisabledFunctionError Traceback (most recent call last) /tmp/ipython-input-709664407.py in <cell line: 0>() 55 56 # Mostrar imagen ---> 57 cv2.imshow("Detections", img) 58 cv2.waitKey(0) 59 cv2.destroyAllWindows() /usr/local/lib/python3.12/dist-packages/google/colab/_import_hooks/_cv2.py in wrapped(*args, **kwargs) 48 def wrapped(*args, **kwargs): 49 if not os.environ.get(env_var, False): ---> 50 raise DisabledFunctionError(message, name or func.__name__) 51 return func(*args, **kwargs) 52 DisabledFunctionError: cv2.imshow() is disabled in Colab, because it causes Jupyter sessions to crash; see https://github.com/jupyter/notebook/issues/3935. As a substitution, consider using from google.colab.patches import cv2_imshow
In [ ]:
import onnxruntime as ort
import cv2
import numpy as np
import matplotlib.pyplot as plt
# -----------------------------
# Config
# -----------------------------
onnx_model_path = "outputs/retinanet.onnx"
image_path = "/content/datasets/african-wildlife/train/images/1 (106).jpg"
input_size = 640 # mismo tamaño que usaste al exportar
CLASSES = ["buffalo", "zebra", "elephant", "rhino"] #s
CONF_THRESH = 0.3 # umbral de confianza
# -----------------------------
# 1️⃣ Cargar y preparar la imagen
# -----------------------------
img = cv2.imread(image_path)
img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img_resized = cv2.resize(img_rgb, (input_size, input_size))
img_normalized = img_resized.astype(np.float32) / 255.0
# Convertir a (1, 3, H, W) para ONNX
input_tensor = np.transpose(img_normalized, (2, 0, 1))[np.newaxis, :, :, :]
# -----------------------------
# 2️⃣ Cargar el modelo ONNX
# -----------------------------
ort_session = ort.InferenceSession(onnx_model_path)
# -----------------------------
# 3️⃣ Inferencia
# -----------------------------
outputs = ort_session.run(None, {"images": input_tensor})
boxes = outputs[0] # shape: [num_detections, 4]
scores = outputs[1] # shape: [num_detections]
labels = outputs[2] # shape: [num_detections]
# -----------------------------
# 4️⃣ Dibujar resultados
# -----------------------------
img_plot = img_rgb.copy()
for box, score, label in zip(boxes, scores, labels):
if score < CONF_THRESH:
continue
x1, y1, x2, y2 = box
# Reescalar a tamaño original
x1 = int(x1 * img.shape[1] / input_size)
x2 = int(x2 * img.shape[1] / input_size)
y1 = int(y1 * img.shape[0] / input_size)
y2 = int(y2 * img.shape[0] / input_size)
class_name = CLASSES[int(label) - 1] # si tus clases empiezan en 1
cv2.rectangle(img_plot, (x1, y1), (x2, y2), (0, 255, 0), 2)
cv2.putText(img_plot, f"{class_name}: {score:.2f}", (x1, max(y1-10, 0)),
cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
# -----------------------------
# 5️⃣ Mostrar imagen en Jupyter/Colab
# -----------------------------
plt.figure(figsize=(12, 8))
plt.imshow(img_plot)
plt.axis('off')
plt.title("Detections ONNX RetinaNet")
plt.show()
In [ ]:
from torchmetrics.detection.mean_ap import MeanAveragePrecision
import torch
metric = MeanAveragePrecision(class_metrics=True) # nos da AP por clase
# Supongamos que tenemos listas de predicciones y targets como en tu entrenamiento
preds = [...] # lista de dicts: {"boxes":..., "scores":..., "labels":...}
targets = [...] # lista de dicts: {"boxes":..., "labels":...}
metric.update(preds, targets)
results = metric.compute()
print(results)
--------------------------------------------------------------------------- TypeError Traceback (most recent call last) /tmp/ipython-input-58285722.py in <cell line: 0>() 8 targets = [...] # lista de dicts: {"boxes":..., "labels":...} 9 ---> 10 metric.update(preds, targets) 11 results = metric.compute() 12 print(results) /usr/local/lib/python3.12/dist-packages/torchmetrics/metric.py in wrapped_func(*args, **kwargs) 547 with torch.set_grad_enabled(self._enable_grad): 548 try: --> 549 update(*args, **kwargs) 550 except RuntimeError as err: 551 if "Expected all tensors to be on" in str(err): /usr/local/lib/python3.12/dist-packages/torchmetrics/detection/mean_ap.py in update(self, preds, target) 528 529 """ --> 530 _input_validator(preds, target, iou_type=self.iou_type) 531 532 for item in preds: /usr/local/lib/python3.12/dist-packages/torchmetrics/detection/helpers.py in _input_validator(preds, targets, iou_type, ignore_score) 64 65 for k in [*item_val_name, "labels"] + (["scores"] if not ignore_score else []): ---> 66 if any(k not in p for p in preds): 67 raise ValueError(f"Expected all dicts in `preds` to contain the `{k}` key") 68 /usr/local/lib/python3.12/dist-packages/torchmetrics/detection/helpers.py in <genexpr>(.0) 64 65 for k in [*item_val_name, "labels"] + (["scores"] if not ignore_score else []): ---> 66 if any(k not in p for p in preds): 67 raise ValueError(f"Expected all dicts in `preds` to contain the `{k}` key") 68 TypeError: argument of type 'ellipsis' is not iterable
In [ ]:
import onnxruntime as ort
import cv2
import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import precision_recall_curve, average_precision_score
# -----------------------------
# Config
# -----------------------------
onnx_model_path = "outputs/retinanet.onnx"
input_size = 640
CLASSES = ["buffalo", "zebra", "elephant", "rhino"] #clases
CONF_THRESH = 0.0 # para PR curve consideramos todas las detecciones
DEVICE = 'cpu' # onnxruntime usa CPU o GPU automáticamente
# -----------------------------
# Cargar ONNX
# -----------------------------
ort_session = ort.InferenceSession(onnx_model_path)
# -----------------------------
# Datos de prueba (ejemplo con 1 imagen)
# En la práctica, recorre todo tu conjunto de validación
# -----------------------------
image_path = "/content/datasets/african-wildlife/val/images/1 (102).jpg"
gt_labels = np.array([1, 2]) # Ejemplo: clases reales presentes (cambia según tu label)
gt_boxes = np.array([[50, 60, 200, 300], [150, 200, 350, 400]]) # xyxy
img = cv2.imread(image_path)
img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img_resized = cv2.resize(img_rgb, (input_size, input_size))
img_normalized = img_resized.astype(np.float32) / 255.0
input_tensor = np.transpose(img_normalized, (2, 0, 1))[np.newaxis, :, :, :]
# -----------------------------
# Inferencia ONNX
# -----------------------------
outputs = ort_session.run(None, {"images": input_tensor})
pred_boxes = outputs[0] # [N,4]
pred_scores = outputs[1] # [N]
pred_labels = outputs[2] # [N]
# -----------------------------
# Flatten para PR curve
# Nota: normalmente usamos todas las imágenes y sus etiquetas
# -----------------------------
y_true = []
y_scores = []
for label in CLASSES:
# Convierte las clases a 0/1 para cada clase
y_true_class = (gt_labels == (CLASSES.index(label)+1)).astype(int)
y_scores_class = pred_scores[(pred_labels == (CLASSES.index(label)+1))]
if len(y_scores_class) == 0:
y_scores_class = np.array([0]*len(y_true_class))
y_true.extend(y_true_class.tolist())
y_scores.extend(y_scores_class.tolist())
# -----------------------------
# Precision-Recall curve
# -----------------------------
precision, recall, thresholds = precision_recall_curve(y_true, y_scores)
ap = average_precision_score(y_true, y_scores)
plt.figure(figsize=(8,6))
plt.plot(recall, precision, marker='.', label=f'AP={ap:.3f}')
plt.xlabel('Recall')
plt.ylabel('Precision')
plt.title('Precision-Recall curve')
plt.legend()
plt.grid(True)
plt.show()
--------------------------------------------------------------------------- ValueError Traceback (most recent call last) /tmp/ipython-input-2268590998.py in <cell line: 0>() 60 # Precision-Recall curve 61 # ----------------------------- ---> 62 precision, recall, thresholds = precision_recall_curve(y_true, y_scores) 63 ap = average_precision_score(y_true, y_scores) 64 /usr/local/lib/python3.12/dist-packages/sklearn/utils/_param_validation.py in wrapper(*args, **kwargs) 214 ) 215 ): --> 216 return func(*args, **kwargs) 217 except InvalidParameterError as e: 218 # When the function is just a wrapper around an estimator, we allow /usr/local/lib/python3.12/dist-packages/sklearn/metrics/_ranking.py in precision_recall_curve(y_true, y_score, pos_label, sample_weight, drop_intermediate, probas_pred) 1003 y_score = probas_pred 1004 -> 1005 fps, tps, thresholds = _binary_clf_curve( 1006 y_true, y_score, pos_label=pos_label, sample_weight=sample_weight 1007 ) /usr/local/lib/python3.12/dist-packages/sklearn/metrics/_ranking.py in _binary_clf_curve(y_true, y_score, pos_label, sample_weight) 818 raise ValueError("{0} format is not supported".format(y_type)) 819 --> 820 check_consistent_length(y_true, y_score, sample_weight) 821 y_true = column_or_1d(y_true) 822 y_score = column_or_1d(y_score) /usr/local/lib/python3.12/dist-packages/sklearn/utils/validation.py in check_consistent_length(*arrays) 473 uniques = np.unique(lengths) 474 if len(uniques) > 1: --> 475 raise ValueError( 476 "Found input variables with inconsistent numbers of samples: %r" 477 % [int(l) for l in lengths] ValueError: Found input variables with inconsistent numbers of samples: [8, 6]
No salió de nuevo, intento cargar el visualization data.py descargado
In [ ]:
!python inf_video.py
usage: inf_video.py [-h] -i INPUT [--imgsz IMGSZ] [--threshold THRESHOLD] inf_video.py: error: the following arguments are required: -i/--input