[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7300MDHSEP7T6B35Q88CEJ2":3,"skill-same:01M7300MDHSEP7T6B35Q88CEJ2":75},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":39,"commitSha":40,"skillDirectory":41,"forksCount":42,"repositoryHtmlUrl":43,"repositoryHomepageUrl":6,"ownerHtmlUrl":44,"ownerBlogUrl":6,"ownerDescription":45,"updatedAt":46,"install":47,"resources":60},"01M7300MDHSEP7T6B35Q88CEJ2","tensorboard",null,"Visualize métricas de treinamento, depure modelos com histogramas, compare experimentos, visualize grafos de modelos e perfil de desempenho com TensorBoard - kit de visualização de ML do Google","artubss\u002FSKILLS-CLAUDE-CODE","artubss","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F205203215?v=4","01M72YSJPDWM2V6S5PB6XP6PM9","Arthur Bezerra ","01M72YSJPZKPKRBGXNG8K6QFA8",12,"MIT","https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fai-research\u002Fmlops-tensorboard\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: tensorboard\ndescription: Visualize métricas de treinamento, depure modelos com histogramas, compare experimentos, visualize grafos de modelos e perfil de desempenho com TensorBoard - kit de visualização de ML do Google\nversion: 1.0.0\nauthor: Orchestra Research\nlicense: MIT\ntags: [MLOps, TensorBoard, Visualization, Training Metrics, Model Debugging, PyTorch, TensorFlow, Experiment Tracking, Performance Profiling]\ndependencies: [tensorboard, torch, tensorflow]\n---\n\n# TensorBoard: Kit de Visualização para ML\n\n## Quando Usar Esta Skill\n\nUse TensorBoard quando você precisa:\n- **Visualizar métricas de treinamento** como perda e acurácia ao longo do tempo\n- **Depurar modelos** com histogramas e distribuições\n- **Comparar experimentos** em múltiplas execuções\n- **Visualizar grafos e arquitetura de modelos**\n- **Projetar embeddings** em dimensões inferiores (t-SNE, PCA)\n- **Rastrear experimentos de hiperparâmetros**\n- **Fazer perfil de desempenho** e identificar gargalos\n- **Visualizar imagens e texto** durante o treinamento\n\n**Usuários**: 20M+ downloads\u002Fano | **GitHub Stars**: 27k+ | **Licença**: Apache 2.0\n\n## Instalação\n\n```bash\n# Instalar TensorBoard\npip install tensorboard\n\n# Integração PyTorch\npip install torch torchvision tensorboard\n\n# Integração TensorFlow (TensorBoard incluído)\npip install tensorflow\n\n# Iniciar TensorBoard\ntensorboard --logdir=runs\n# Acesse em http:\u002F\u002Flocalhost:6006\n```\n\n## Quick Start\n\n### PyTorch\n\n```python\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Criar writer\nwriter = SummaryWriter('runs\u002Fexperiment_1')\n\n# Loop de treinamento\nfor epoch in range(10):\n    train_loss = train_epoch()\n    val_acc = validate()\n\n    # Registrar métricas\n    writer.add_scalar('Loss\u002Ftrain', train_loss, epoch)\n    writer.add_scalar('Accuracy\u002Fval', val_acc, epoch)\n\n# Fechar writer\nwriter.close()\n\n# Iniciar: tensorboard --logdir=runs\n```\n\n### TensorFlow\u002FKeras\n\n```python\nimport tensorflow as tf\n\n# Criar callback\ntensorboard_callback = tf.keras.callbacks.TensorBoard(\n    log_dir='logs\u002Ffit',\n    histogram_freq=1\n)\n\n# Treinar modelo\nmodel.fit(\n    x_train, y_train,\n    epochs=10,\n    validation_data=(x_val, y_val),\n    callbacks=[tensorboard_callback]\n)\n\n# Iniciar: tensorboard --logdir=logs\n```\n\n## Conceitos Principais\n\n### 1. SummaryWriter (PyTorch)\n\n```python\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Diretório padrão: runs\u002FCURRENT_DATETIME\nwriter = SummaryWriter()\n\n# Diretório personalizado\nwriter = SummaryWriter('runs\u002Fexperiment_1')\n\n# Comentário personalizado (anexado ao diretório padrão)\nwriter = SummaryWriter(comment='baseline')\n\n# Registrar dados\nwriter.add_scalar('Loss\u002Ftrain', 0.5, step=0)\nwriter.add_scalar('Loss\u002Ftrain', 0.3, step=1)\n\n# Descarregar e fechar\nwriter.flush()\nwriter.close()\n```\n\n### 2. Logging de Escalares\n\n```python\n# PyTorch\nfrom torch.utils.tensorboard import SummaryWriter\nwriter = SummaryWriter()\n\nfor epoch in range(100):\n    train_loss = train()\n    val_loss = validate()\n\n    # Registrar métricas individuais\n    writer.add_scalar('Loss\u002Ftrain', train_loss, epoch)\n    writer.add_scalar('Loss\u002Fval', val_loss, epoch)\n    writer.add_scalar('Accuracy\u002Ftrain', train_acc, epoch)\n    writer.add_scalar('Accuracy\u002Fval', val_acc, epoch)\n\n    # Taxa de aprendizado\n    lr = optimizer.param_groups[0]['lr']\n    writer.add_scalar('Learning_rate', lr, epoch)\n\nwriter.close()\n```\n\n```python\n# TensorFlow\nimport tensorflow as tf\n\ntrain_summary_writer = tf.summary.create_file_writer('logs\u002Ftrain')\nval_summary_writer = tf.summary.create_file_writer('logs\u002Fval')\n\nfor epoch in range(100):\n    with train_summary_writer.as_default():\n        tf.summary.scalar('loss', train_loss, step=epoch)\n        tf.summary.scalar('accuracy', train_acc, step=epoch)\n\n    with val_summary_writer.as_default():\n        tf.summary.scalar('loss', val_loss, step=epoch)\n        tf.summary.scalar('accuracy', val_acc, step=epoch)\n```\n\n### 3. Logging de Múltiplos Escalares\n\n```python\n# PyTorch: Agrupar métricas relacionadas\nwriter.add_scalars('Loss', {\n    'train': train_loss,\n    'validation': val_loss,\n    'test': test_loss\n}, epoch)\n\nwriter.add_scalars('Metrics', {\n    'accuracy': accuracy,\n    'precision': precision,\n    'recall': recall,\n    'f1': f1_score\n}, epoch)\n```\n\n### 4. Logging de Imagens\n\n```python\n# PyTorch\nimport torch\nfrom torchvision.utils import make_grid\n\n# Imagem única\nwriter.add_image('Input\u002Fsample', img_tensor, epoch)\n\n# Múltiplas imagens em grade\nimg_grid = make_grid(images[:64], nrow=8)\nwriter.add_image('Batch\u002Finputs', img_grid, epoch)\n\n# Visualização de predições\npred_grid = make_grid(predictions[:16], nrow=4)\nwriter.add_image('Predictions', pred_grid, epoch)\n```\n\n```python\n# TensorFlow\nimport tensorflow as tf\n\nwith file_writer.as_default():\n    # Codificar imagens como PNG\n    tf.summary.image('Training samples', images, step=epoch, max_outputs=25)\n```\n\n### 5. Logging de Histogramas\n\n```python\n# PyTorch: Rastrear distribuições de pesos\nfor name, param in model.named_parameters():\n    writer.add_histogram(name, param, epoch)\n\n    # Rastrear gradientes\n    if param.grad is not None:\n        writer.add_histogram(f'{name}.grad', param.grad, epoch)\n\n# Rastrear ativações\nwriter.add_histogram('Activations\u002Frelu1', activations, epoch)\n```\n\n```python\n# TensorFlow\nwith file_writer.as_default():\n    tf.summary.histogram('weights\u002Flayer1', layer1.kernel, step=epoch)\n    tf.summary.histogram('activations\u002Frelu1', activations, step=epoch)\n```\n\n### 6. Logging de Grafo de Modelo\n\n```python\n# PyTorch\nimport torch\n\nmodel = MyModel()\ndummy_input = torch.randn(1, 3, 224, 224)\n\nwriter.add_graph(model, dummy_input)\nwriter.close()\n```\n\n```python\n# TensorFlow (automático com Keras)\ntensorboard_callback = tf.keras.callbacks.TensorBoard(\n    log_dir='logs',\n    write_graph=True\n)\n\nmodel.fit(x, y, callbacks=[tensorboard_callback])\n```\n\n## Recursos Avançados\n\n### Embedding Projector\n\nVisualize dados de alta dimensionalidade (embeddings, features) em 2D\u002F3D.\n\n```python\nimport torch\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Obter embeddings (ex: embeddings de palavras, features de imagem)\nembeddings = model.get_embeddings(data)  # Shape: (N, embedding_dim)\n\n# Metadados (rótulos para cada ponto)\nmetadata = ['class_1', 'class_2', 'class_1', ...]\n\n# Imagens (opcional, para embeddings de imagem)\nlabel_images = torch.stack([img1, img2, img3, ...])\n\n# Registrar em TensorBoard\nwriter.add_embedding(\n    embeddings,\n    metadata=metadata,\n    label_img=label_images,\n    global_step=epoch\n)\n```\n\n**Em TensorBoard:**\n- Navegue até a aba \"Projector\"\n- Escolha visualização PCA, t-SNE ou UMAP\n- Pesquise, filtre e explore clusters\n\n### Ajuste de Hiperparâmetros\n\n```python\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Experimentar diferentes hiperparâmetros\nfor lr in [0.001, 0.01, 0.1]:\n    for batch_size in [16, 32, 64]:\n        # Criar diretório de execução único\n        writer = SummaryWriter(f'runs\u002Flr{lr}_bs{batch_size}')\n\n        # Registrar hiperparâmetros\n        writer.add_hparams(\n            {'lr': lr, 'batch_size': batch_size},\n            {'hparam\u002Faccuracy': final_acc, 'hparam\u002Floss': final_loss}\n        )\n\n        # Treinar e registrar\n        for epoch in range(10):\n            loss = train(lr, batch_size)\n            writer.add_scalar('Loss\u002Ftrain', loss, epoch)\n\n        writer.close()\n\n# Comparar na aba \"HParams\" do TensorBoard\n```\n\n### Logging de Texto\n\n```python\n# PyTorch: Registrar texto (ex: predições de modelo, resumos)\nwriter.add_text('Predictions', f'Epoch {epoch}: {predictions}', epoch)\nwriter.add_text('Config', str(config), 0)\n\n# Registrar tabelas markdown\nmarkdown_table = \"\"\"\n| Metric | Value |\n|--------|-------|\n| Accuracy | 0.95 |\n| F1 Score | 0.93 |\n\"\"\"\nwriter.add_text('Results', markdown_table, epoch)\n```\n\n### Curvas PR\n\nCurvas de Precisão-Recall para classificação.\n\n```python\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Obter predições e rótulos\npredictions = model(test_data)  # Shape: (N, num_classes)\nlabels = test_labels  # Shape: (N,)\n\n# Registrar curva PR para cada classe\nfor i in range(num_classes):\n    writer.add_pr_curve(\n        f'PR_curve\u002Fclass_{i}',\n        labels == i,\n        predictions[:, i],\n        global_step=epoch\n    )\n```\n\n## Exemplos de Integração\n\n### Loop de Treinamento PyTorch\n\n```python\nimport torch\nimport torch.nn as nn\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Setup\nwriter = SummaryWriter('runs\u002Fresnet_experiment')\nmodel = ResNet50()\noptimizer = torch.optim.Adam(model.parameters(), lr=0.001)\ncriterion = nn.CrossEntropyLoss()\n\n# Registrar grafo do modelo\ndummy_input = torch.randn(1, 3, 224, 224)\nwriter.add_graph(model, dummy_input)\n\n# Loop de treinamento\nfor epoch in range(50):\n    model.train()\n    train_loss = 0.0\n    train_correct = 0\n\n    for batch_idx, (data, target) in enumerate(train_loader):\n        optimizer.zero_grad()\n        output = model(data)\n        loss = criterion(output, target)\n        loss.backward()\n        optimizer.step()\n\n        train_loss += loss.item()\n        pred = output.argmax(dim=1)\n        train_correct += pred.eq(target).sum().item()\n\n        # Registrar métricas de batch (a cada 100 batches)\n        if batch_idx % 100 == 0:\n            global_step = epoch * len(train_loader) + batch_idx\n            writer.add_scalar('Loss\u002Ftrain_batch', loss.item(), global_step)\n\n    # Métricas de epoch\n    train_loss \u002F= len(train_loader)\n    train_acc = train_correct \u002F len(train_loader.dataset)\n\n    # Validação\n    model.eval()\n    val_loss = 0.0\n    val_correct = 0\n\n    with torch.no_grad():\n        for data, target in val_loader:\n            output = model(data)\n            val_loss += criterion(output, target).item()\n            pred = output.argmax(dim=1)\n            val_correct += pred.eq(target).sum().item()\n\n    val_loss \u002F= len(val_loader)\n    val_acc = val_correct \u002F len(val_loader.dataset)\n\n    # Registrar métricas de epoch\n    writer.add_scalars('Loss', {'train': train_loss, 'val': val_loss}, epoch)\n    writer.add_scalars('Accuracy', {'train': train_acc, 'val': val_acc}, epoch)\n\n    # Registrar taxa de aprendizado\n    writer.add_scalar('Learning_rate', optimizer.param_groups[0]['lr'], epoch)\n\n    # Registrar histogramas (a cada 5 epochs)\n    if epoch % 5 == 0:\n        for name, param in model.named_parameters():\n            writer.add_histogram(name, param, epoch)\n\n    # Registrar predições de amostra\n    if epoch % 10 == 0:\n        sample_images = data[:8]\n        writer.add_image('Sample_inputs', make_grid(sample_images), epoch)\n\nwriter.close()\n```\n\n### Treinamento TensorFlow\u002FKeras\n\n```python\nimport tensorflow as tf\n\n# Definir modelo\nmodel = tf.keras.models.Sequential([\n    tf.keras.layers.Conv2D(32, 3, activation='relu', input_shape=(28, 28, 1)),\n    tf.keras.layers.MaxPooling2D(),\n    tf.keras.layers.Flatten(),\n    tf.keras.layers.Dense(128, activation='relu'),\n    tf.keras.layers.Dense(10, activation='softmax')\n])\n\nmodel.compile(\n    optimizer='adam',\n    loss='sparse_categorical_crossentropy',\n    metrics=['accuracy']\n)\n\n# Callback do TensorBoard\ntensorboard_callback = tf.keras.callbacks.TensorBoard(\n    log_dir='logs\u002Ffit',\n    histogram_freq=1,          # Registrar histogramas a cada epoch\n    write_graph=True,          # Visualizar grafo do modelo\n    write_images=True,         # Visualizar pesos como imagens\n    update_freq='epoch',       # Registrar métricas a cada epoch\n    profile_batch='500,520',   # Fazer perfil dos batches 500-520\n    embeddings_freq=1          # Registrar embeddings a cada epoch\n)\n\n# Treinar\nmodel.fit(\n    x_train, y_train,\n    epochs=10,\n    validation_data=(x_val, y_val),\n    callbacks=[tensorboard_callback]\n)\n```\n\n## Comparando Experimentos\n\n### Múltiplas Execuções\n\n```bash\n# Executar experimentos com diferentes configs\npython train.py --lr 0.001 --logdir runs\u002Fexp1\npython train.py --lr 0.01 --logdir runs\u002Fexp2\npython train.py --lr 0.1 --logdir runs\u002Fexp3\n\n# Visualizar todas as execuções junto\ntensorboard --logdir=runs\n```\n\n**Em TensorBoard:**\n- Todas as execuções aparecem no mesmo dashboard\n- Ativar\u002Fdesativar execuções para comparação\n- Usar regex para filtrar nomes de execução\n- Sobrepor gráficos para comparar métricas\n\n### Organizando Experimentos\n\n```python\n# Organização hierárquica\nruns\u002F\n├── baseline\u002F\n│   ├── run_1\u002F\n│   └── run_2\u002F\n├── improved\u002F\n│   ├── run_1\u002F\n│   └── run_2\u002F\n└── final\u002F\n    └── run_1\u002F\n\n# Registrar com hierarquia\nwriter = SummaryWriter('runs\u002Fbaseline\u002Frun_1')\n```\n\n## Melhores Práticas\n\n### 1. Use Nomes Descritivos para Execuções\n\n```python\n# ✅ Bom: Nomes descritivos\nfrom datetime import datetime\ntimestamp = datetime.now().strftime('%Y%m%d_%H%M%S')\nwriter = SummaryWriter(f'runs\u002Fresnet50_lr0.001_bs32_{timestamp}')\n\n# ❌ Ruim: Nomes auto-gerados\nwriter = SummaryWriter()  # Cria runs\u002FJan01_12-34-56_hostname\n```\n\n### 2. Agrupar Métricas Relacionadas\n\n```python\n# ✅ Bom: Métricas agrupadas\nwriter.add_scalar('Loss\u002Ftrain', train_loss, step)\nwriter.add_scalar('Loss\u002Fval', val_loss, step)\nwriter.add_scalar('Accuracy\u002Ftrain', train_acc, step)\nwriter.add_scalar('Accuracy\u002Fval', val_acc, step)\n\n# ❌ Ruim: Namespace plano\nwriter.add_scalar('train_loss', train_loss, step)\nwriter.add_scalar('val_loss', val_loss, step)\n```\n\n### 3. Registre Regularmente Mas Não Muito Frequentemente\n\n```python\n# ✅ Bom: Sempre registrar métricas de epoch, ocasionalmente de batch\nfor epoch in range(100):\n    for batch_idx, (data, target) in enumerate(train_loader):\n        loss = train_step(data, target)\n\n        # Registrar a cada 100 batches\n        if batch_idx % 100 == 0:\n            writer.add_scalar('Loss\u002Fbatch', loss, global_step)\n\n    # Sempre registrar métricas de epoch\n    writer.add_scalar('Loss\u002Fepoch', epoch_loss, epoch)\n\n# ❌ Ruim: Registrar cada batch (cria arquivos de log enormes)\nfor batch in train_loader:\n    writer.add_scalar('Loss', loss, step)  # Muito frequente\n```\n\n### 4. Feche o Writer Quando Terminar\n\n```python\n# ✅ Bom: Usar gerenciador de contexto\nwith SummaryWriter('runs\u002Fexp1') as writer:\n    for epoch in range(10):\n        writer.add_scalar('Loss', loss, epoch)\n# Fecha automaticamente\n\n# Ou manualmente\nwriter = SummaryWriter('runs\u002Fexp1')\n# ... registrar ...\nwriter.close()\n```\n\n### 5. Use Writers Separados para Treinamento\u002FValidação\n\n```python\n# ✅ Bom: Diretórios de log separados\ntrain_writer = SummaryWriter('runs\u002Fexp1\u002Ftrain')\nval_writer = SummaryWriter('runs\u002Fexp1\u002Fval')\n\ntrain_writer.add_scalar('loss', train_loss, epoch)\nval_writer.add_scalar('loss', val_loss, epoch)\n```\n\n## Perfil de Desempenho\n\n### TensorFlow Profiler\n\n```python\n# Ativar profiling\ntensorboard_callback = tf.keras.callbacks.TensorBoard(\n    log_dir='logs',\n    profile_batch='10,20'  # Fazer perfil dos batches 10-20\n)\n\nmodel.fit(x, y, callbacks=[tensorboard_callback])\n\n# Visualizar na aba Profile do TensorBoard\n# Mostra: utilização de GPU, estatísticas de kernel, uso de memória, gargalos\n```\n\n### PyTorch Profiler\n\n```python\nimport torch.profiler as profiler\n\nwith profiler.profile(\n    activities=[\n        profiler.ProfilerActivity.CPU,\n        profiler.ProfilerActivity.CUDA\n    ],\n    on_trace_ready=torch.profiler.tensorboard_trace_handler('.\u002Fruns\u002Fprofiler'),\n    record_shapes=True,\n    with_stack=True\n) as prof:\n    for batch in train_loader:\n        loss = train_step(batch)\n        prof.step()\n\n# Visualizar na aba Profile do TensorBoard\n```\n\n## Recursos\n\n- **Documentação**: https:\u002F\u002Fwww.tensorflow.org\u002Ftensorboard\n- **Integração PyTorch**: https:\u002F\u002Fpytorch.org\u002Fdocs\u002Fstable\u002Ftensorboard.html\n- **GitHub**: https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Ftensorboard (27k+ stars)\n- **TensorBoard.dev**: https:\u002F\u002Ftensorboard.dev (compartilhar experimentos publicamente)\n\n## Veja Também\n\n- `references\u002Fvisualization.md` - Guia abrangente de visualização\n- `references\u002Fprofiling.md` - Padrões de perfil de desempenho\n- `references\u002Fintegrations.md` - Exemplos de integração específicos da framework","# TensorBoard: Kit de Visualização para ML\n\n## Quando Usar Esta Skill\n\nUse TensorBoard quando você precisa:\n- **Visualizar métricas de treinamento** como perda e acurácia ao longo do tempo\n- **Depurar modelos** com histogramas e distribuições\n- **Comparar experimentos** em múltiplas execuções\n- **Visualizar grafos e arquitetura de modelos**\n- **Projetar embeddings** em dimensões inferiores (t-SNE, PCA)\n- **Rastrear experimentos de hiperparâmetros**\n- **Fazer perfil de desempenho** e identificar gargalos\n- **Visualizar imagens e texto** durante o treinamento\n\n**Usuários**: 20M+ downloads\u002Fano | **GitHub Stars**: 27k+ | **Licença**: Apache 2.0\n\n## Instalação\n\n```bash\n# Instalar TensorBoard\npip install tensorboard\n\n# Integração PyTorch\npip install torch torchvision tensorboard\n\n# Integração TensorFlow (TensorBoard incluído)\npip install tensorflow\n\n# Iniciar TensorBoard\ntensorboard --logdir=runs\n# Acesse em http:\u002F\u002Flocalhost:6006\n```\n\n## Quick Start\n\n### PyTorch\n\n```python\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Criar writer\nwriter = SummaryWriter('runs\u002Fexperiment_1')\n\n# Loop de treinamento\nfor epoch in range(10):\n    train_loss = train_epoch()\n    val_acc = validate()\n\n    # Registrar métricas\n    writer.add_scalar('Loss\u002Ftrain', train_loss, epoch)\n    writer.add_scalar('Accuracy\u002Fval', val_acc, epoch)\n\n# Fechar writer\nwriter.close()\n\n# Iniciar: tensorboard --logdir=runs\n```\n\n### TensorFlow\u002FKeras\n\n```python\nimport tensorflow as tf\n\n# Criar callback\ntensorboard_callback = tf.keras.callbacks.TensorBoard(\n    log_dir='logs\u002Ffit',\n    histogram_freq=1\n)\n\n# Treinar modelo\nmodel.fit(\n    x_train, y_train,\n    epochs=10,\n    validation_data=(x_val, y_val),\n    callbacks=[tensorboard_callback]\n)\n\n# Iniciar: tensorboard --logdir=logs\n```\n\n## Conceitos Principais\n\n### 1. SummaryWriter (PyTorch)\n\n```python\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Diretório padrão: runs\u002FCURRENT_DATETIME\nwriter = SummaryWriter()\n\n# Diretório personalizado\nwriter = SummaryWriter('runs\u002Fexperiment_1')\n\n# Comentário personalizado (anexado ao diretório padrão)\nwriter = SummaryWriter(comment='baseline')\n\n# Registrar dados\nwriter.add_scalar('Loss\u002Ftrain', 0.5, step=0)\nwriter.add_scalar('Loss\u002Ftrain', 0.3, step=1)\n\n# Descarregar e fechar\nwriter.flush()\nwriter.close()\n```\n\n### 2. Logging de Escalares\n\n```python\n# PyTorch\nfrom torch.utils.tensorboard import SummaryWriter\nwriter = SummaryWriter()\n\nfor epoch in range(100):\n    train_loss = train()\n    val_loss = validate()\n\n    # Registrar métricas individuais\n    writer.add_scalar('Loss\u002Ftrain', train_loss, epoch)\n    writer.add_scalar('Loss\u002Fval', val_loss, epoch)\n    writer.add_scalar('Accuracy\u002Ftrain', train_acc, epoch)\n    writer.add_scalar('Accuracy\u002Fval', val_acc, epoch)\n\n    # Taxa de aprendizado\n    lr = optimizer.param_groups[0]['lr']\n    writer.add_scalar('Learning_rate', lr, epoch)\n\nwriter.close()\n```\n\n```python\n# TensorFlow\nimport tensorflow as tf\n\ntrain_summary_writer = tf.summary.create_file_writer('logs\u002Ftrain')\nval_summary_writer = tf.summary.create_file_writer('logs\u002Fval')\n\nfor epoch in range(100):\n    with train_summary_writer.as_default():\n        tf.summary.scalar('loss', train_loss, step=epoch)\n        tf.summary.scalar('accuracy', train_acc, step=epoch)\n\n    with val_summary_writer.as_default():\n        tf.summary.scalar('loss', val_loss, step=epoch)\n        tf.summary.scalar('accuracy', val_acc, step=epoch)\n```\n\n### 3. Logging de Múltiplos Escalares\n\n```python\n# PyTorch: Agrupar métricas relacionadas\nwriter.add_scalars('Loss', {\n    'train': train_loss,\n    'validation': val_loss,\n    'test': test_loss\n}, epoch)\n\nwriter.add_scalars('Metrics', {\n    'accuracy': accuracy,\n    'precision': precision,\n    'recall': recall,\n    'f1': f1_score\n}, epoch)\n```\n\n### 4. Logging de Imagens\n\n```python\n# PyTorch\nimport torch\nfrom torchvision.utils import make_grid\n\n# Imagem única\nwriter.add_image('Input\u002Fsample', img_tensor, epoch)\n\n# Múltiplas imagens em grade\nimg_grid = make_grid(images[:64], nrow=8)\nwriter.add_image('Batch\u002Finputs', img_grid, epoch)\n\n# Visualização de predições\npred_grid = make_grid(predictions[:16], nrow=4)\nwriter.add_image('Predictions', pred_grid, epoch)\n```\n\n```python\n# TensorFlow\nimport tensorflow as tf\n\nwith file_writer.as_default():\n    # Codificar imagens como PNG\n    tf.summary.image('Training samples', images, step=epoch, max_outputs=25)\n```\n\n### 5. Logging de Histogramas\n\n```python\n# PyTorch: Rastrear distribuições de pesos\nfor name, param in model.named_parameters():\n    writer.add_histogram(name, param, epoch)\n\n    # Rastrear gradientes\n    if param.grad is not None:\n        writer.add_histogram(f'{name}.grad', param.grad, epoch)\n\n# Rastrear ativações\nwriter.add_histogram('Activations\u002Frelu1', activations, epoch)\n```\n\n```python\n# TensorFlow\nwith file_writer.as_default():\n    tf.summary.histogram('weights\u002Flayer1', layer1.kernel, step=epoch)\n    tf.summary.histogram('activations\u002Frelu1', activations, step=epoch)\n```\n\n### 6. Logging de Grafo de Modelo\n\n```python\n# PyTorch\nimport torch\n\nmodel = MyModel()\ndummy_input = torch.randn(1, 3, 224, 224)\n\nwriter.add_graph(model, dummy_input)\nwriter.close()\n```\n\n```python\n# TensorFlow (automático com Keras)\ntensorboard_callback = tf.keras.callbacks.TensorBoard(\n    log_dir='logs',\n    write_graph=True\n)\n\nmodel.fit(x, y, callbacks=[tensorboard_callback])\n```\n\n## Recursos Avançados\n\n### Embedding Projector\n\nVisualize dados de alta dimensionalidade (embeddings, features) em 2D\u002F3D.\n\n```python\nimport torch\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Obter embeddings (ex: embeddings de palavras, features de imagem)\nembeddings = model.get_embeddings(data)  # Shape: (N, embedding_dim)\n\n# Metadados (rótulos para cada ponto)\nmetadata = ['class_1', 'class_2', 'class_1', ...]\n\n# Imagens (opcional, para embeddings de imagem)\nlabel_images = torch.stack([img1, img2, img3, ...])\n\n# Registrar em TensorBoard\nwriter.add_embedding(\n    embeddings,\n    metadata=metadata,\n    label_img=label_images,\n    global_step=epoch\n)\n```\n\n**Em TensorBoard:**\n- Navegue até a aba \"Projector\"\n- Escolha visualização PCA, t-SNE ou UMAP\n- Pesquise, filtre e explore clusters\n\n### Ajuste de Hiperparâmetros\n\n```python\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Experimentar diferentes hiperparâmetros\nfor lr in [0.001, 0.01, 0.1]:\n    for batch_size in [16, 32, 64]:\n        # Criar diretório de execução único\n        writer = SummaryWriter(f'runs\u002Flr{lr}_bs{batch_size}')\n\n        # Registrar hiperparâmetros\n        writer.add_hparams(\n            {'lr': lr, 'batch_size': batch_size},\n            {'hparam\u002Faccuracy': final_acc, 'hparam\u002Floss': final_loss}\n        )\n\n        # Treinar e registrar\n        for epoch in range(10):\n            loss = train(lr, batch_size)\n            writer.add_scalar('Loss\u002Ftrain', loss, epoch)\n\n        writer.close()\n\n# Comparar na aba \"HParams\" do TensorBoard\n```\n\n### Logging de Texto\n\n```python\n# PyTorch: Registrar texto (ex: predições de modelo, resumos)\nwriter.add_text('Predictions', f'Epoch {epoch}: {predictions}', epoch)\nwriter.add_text('Config', str(config), 0)\n\n# Registrar tabelas markdown\nmarkdown_table = \"\"\"\n| Metric | Value |\n|--------|-------|\n| Accuracy | 0.95 |\n| F1 Score | 0.93 |\n\"\"\"\nwriter.add_text('Results', markdown_table, epoch)\n```\n\n### Curvas PR\n\nCurvas de Precisão-Recall para classificação.\n\n```python\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Obter predições e rótulos\npredictions = model(test_data)  # Shape: (N, num_classes)\nlabels = test_labels  # Shape: (N,)\n\n# Registrar curva PR para cada classe\nfor i in range(num_classes):\n    writer.add_pr_curve(\n        f'PR_curve\u002Fclass_{i}',\n        labels == i,\n        predictions[:, i],\n        global_step=epoch\n    )\n```\n\n## Exemplos de Integração\n\n### Loop de Treinamento PyTorch\n\n```python\nimport torch\nimport torch.nn as nn\nfrom torch.utils.tensorboard import SummaryWriter\n\n# Setup\nwriter = SummaryWriter('runs\u002Fresnet_experiment')\nmodel = ResNet50()\noptimizer = torch.optim.Adam(model.parameters(), lr=0.001)\ncriterion = nn.CrossEntropyLoss()\n\n# Registrar grafo do modelo\ndummy_input = torch.randn(1, 3, 224, 224)\nwriter.add_graph(model, dummy_input)\n\n# Loop de treinamento\nfor epoch in range(50):\n    model.train()\n    train_loss = 0.0\n    train_correct = 0\n\n    for batch_idx, (data, target) in enumerate(train_loader):\n        optimizer.zero_grad()\n        output = model(data)\n        loss = criterion(output, target)\n        loss.backward()\n        optimizer.step()\n\n        train_loss += loss.item()\n        pred = output.argmax(dim=1)\n        train_correct += pred.eq(target).sum().item()\n\n        # Registrar métricas de batch (a cada 100 batches)\n        if batch_idx % 100 == 0:\n            global_step = epoch * len(train_loader) + batch_idx\n            writer.add_scalar('Loss\u002Ftrain_batch', loss.item(), global_step)\n\n    # Métricas de epoch\n    train_loss \u002F= len(train_loader)\n    train_acc = train_correct \u002F len(train_loader.dataset)\n\n    # Validação\n    model.eval()\n    val_loss = 0.0\n    val_correct = 0\n\n    with torch.no_grad():\n        for data, target in val_loader:\n            output = model(data)\n            val_loss += criterion(output, target).item()\n            pred = output.argmax(dim=1)\n            val_correct += pred.eq(target).sum().item()\n\n    val_loss \u002F= len(val_loader)\n    val_acc = val_correct \u002F len(val_loader.dataset)\n\n    # Registrar métricas de epoch\n    writer.add_scalars('Loss', {'train': train_loss, 'val': val_loss}, epoch)\n    writer.add_scalars('Accuracy', {'train': train_acc, 'val': val_acc}, epoch)\n\n    # Registrar taxa de aprendizado\n    writer.add_scalar('Learning_rate', optimizer.param_groups[0]['lr'], epoch)\n\n    # Registrar histogramas (a cada 5 epochs)\n    if epoch % 5 == 0:\n        for name, param in model.named_parameters():\n            writer.add_histogram(name, param, epoch)\n\n    # Registrar predições de amostra\n    if epoch % 10 == 0:\n        sample_images = data[:8]\n        writer.add_image('Sample_inputs', make_grid(sample_images), epoch)\n\nwriter.close()\n```\n\n### Treinamento TensorFlow\u002FKeras\n\n```python\nimport tensorflow as tf\n\n# Definir modelo\nmodel = tf.keras.models.Sequential([\n    tf.keras.layers.Conv2D(32, 3, activation='relu', input_shape=(28, 28, 1)),\n    tf.keras.layers.MaxPooling2D(),\n    tf.keras.layers.Flatten(),\n    tf.keras.layers.Dense(128, activation='relu'),\n    tf.keras.layers.Dense(10, activation='softmax')\n])\n\nmodel.compile(\n    optimizer='adam',\n    loss='sparse_categorical_crossentropy',\n    metrics=['accuracy']\n)\n\n# Callback do TensorBoard\ntensorboard_callback = tf.keras.callbacks.TensorBoard(\n    log_dir='logs\u002Ffit',\n    histogram_freq=1,          # Registrar histogramas a cada epoch\n    write_graph=True,          # Visualizar grafo do modelo\n    write_images=True,         # Visualizar pesos como imagens\n    update_freq='epoch',       # Registrar métricas a cada epoch\n    profile_batch='500,520',   # Fazer perfil dos batches 500-520\n    embeddings_freq=1          # Registrar embeddings a cada epoch\n)\n\n# Treinar\nmodel.fit(\n    x_train, y_train,\n    epochs=10,\n    validation_data=(x_val, y_val),\n    callbacks=[tensorboard_callback]\n)\n```\n\n## Comparando Experimentos\n\n### Múltiplas Execuções\n\n```bash\n# Executar experimentos com diferentes configs\npython train.py --lr 0.001 --logdir runs\u002Fexp1\npython train.py --lr 0.01 --logdir runs\u002Fexp2\npython train.py --lr 0.1 --logdir runs\u002Fexp3\n\n# Visualizar todas as execuções junto\ntensorboard --logdir=runs\n```\n\n**Em TensorBoard:**\n- Todas as execuções aparecem no mesmo dashboard\n- Ativar\u002Fdesativar execuções para comparação\n- Usar regex para filtrar nomes de execução\n- Sobrepor gráficos para comparar métricas\n\n### Organizando Experimentos\n\n```python\n# Organização hierárquica\nruns\u002F\n├── baseline\u002F\n│   ├── run_1\u002F\n│   └── run_2\u002F\n├── improved\u002F\n│   ├── run_1\u002F\n│   └── run_2\u002F\n└── final\u002F\n    └── run_1\u002F\n\n# Registrar com hierarquia\nwriter = SummaryWriter('runs\u002Fbaseline\u002Frun_1')\n```\n\n## Melhores Práticas\n\n### 1. Use Nomes Descritivos para Execuções\n\n```python\n# ✅ Bom: Nomes descritivos\nfrom datetime import datetime\ntimestamp = datetime.now().strftime('%Y%m%d_%H%M%S')\nwriter = SummaryWriter(f'runs\u002Fresnet50_lr0.001_bs32_{timestamp}')\n\n# ❌ Ruim: Nomes auto-gerados\nwriter = SummaryWriter()  # Cria runs\u002FJan01_12-34-56_hostname\n```\n\n### 2. Agrupar Métricas Relacionadas\n\n```python\n# ✅ Bom: Métricas agrupadas\nwriter.add_scalar('Loss\u002Ftrain', train_loss, step)\nwriter.add_scalar('Loss\u002Fval', val_loss, step)\nwriter.add_scalar('Accuracy\u002Ftrain', train_acc, step)\nwriter.add_scalar('Accuracy\u002Fval', val_acc, step)\n\n# ❌ Ruim: Namespace plano\nwriter.add_scalar('train_loss', train_loss, step)\nwriter.add_scalar('val_loss', val_loss, step)\n```\n\n### 3. Registre Regularmente Mas Não Muito Frequentemente\n\n```python\n# ✅ Bom: Sempre registrar métricas de epoch, ocasionalmente de batch\nfor epoch in range(100):\n    for batch_idx, (data, target) in enumerate(train_loader):\n        loss = train_step(data, target)\n\n        # Registrar a cada 100 batches\n        if batch_idx % 100 == 0:\n            writer.add_scalar('Loss\u002Fbatch', loss, global_step)\n\n    # Sempre registrar métricas de epoch\n    writer.add_scalar('Loss\u002Fepoch', epoch_loss, epoch)\n\n# ❌ Ruim: Registrar cada batch (cria arquivos de log enormes)\nfor batch in train_loader:\n    writer.add_scalar('Loss', loss, step)  # Muito frequente\n```\n\n### 4. Feche o Writer Quando Terminar\n\n```python\n# ✅ Bom: Usar gerenciador de contexto\nwith SummaryWriter('runs\u002Fexp1') as writer:\n    for epoch in range(10):\n        writer.add_scalar('Loss', loss, epoch)\n# Fecha automaticamente\n\n# Ou manualmente\nwriter = SummaryWriter('runs\u002Fexp1')\n# ... registrar ...\nwriter.close()\n```\n\n### 5. Use Writers Separados para Treinamento\u002FValidação\n\n```python\n# ✅ Bom: Diretórios de log separados\ntrain_writer = SummaryWriter('runs\u002Fexp1\u002Ftrain')\nval_writer = SummaryWriter('runs\u002Fexp1\u002Fval')\n\ntrain_writer.add_scalar('loss', train_loss, epoch)\nval_writer.add_scalar('loss', val_loss, epoch)\n```\n\n## Perfil de Desempenho\n\n### TensorFlow Profiler\n\n```python\n# Ativar profiling\ntensorboard_callback = tf.keras.callbacks.TensorBoard(\n    log_dir='logs',\n    profile_batch='10,20'  # Fazer perfil dos batches 10-20\n)\n\nmodel.fit(x, y, callbacks=[tensorboard_callback])\n\n# Visualizar na aba Profile do TensorBoard\n# Mostra: utilização de GPU, estatísticas de kernel, uso de memória, gargalos\n```\n\n### PyTorch Profiler\n\n```python\nimport torch.profiler as profiler\n\nwith profiler.profile(\n    activities=[\n        profiler.ProfilerActivity.CPU,\n        profiler.ProfilerActivity.CUDA\n    ],\n    on_trace_ready=torch.profiler.tensorboard_trace_handler('.\u002Fruns\u002Fprofiler'),\n    record_shapes=True,\n    with_stack=True\n) as prof:\n    for batch in train_loader:\n        loss = train_step(batch)\n        prof.step()\n\n# Visualizar na aba Profile do TensorBoard\n```\n\n## Recursos\n\n- **Documentação**: https:\u002F\u002Fwww.tensorflow.org\u002Ftensorboard\n- **Integração PyTorch**: https:\u002F\u002Fpytorch.org\u002Fdocs\u002Fstable\u002Ftensorboard.html\n- **GitHub**: https:\u002F\u002Fgithub.com\u002Ftensorflow\u002Ftensorboard (27k+ stars)\n- **TensorBoard.dev**: https:\u002F\u002Ftensorboard.dev (compartilhar experimentos publicamente)\n\n## Veja Também\n\n- `references\u002Fvisualization.md` - Guia abrangente de visualização\n- `references\u002Fprofiling.md` - Padrões de perfil de desempenho\n- `references\u002Fintegrations.md` - Exemplos de integração específicos da framework",{"name":5,"tags":24,"author":34,"license":15,"version":35,"description":7,"dependencies":36},[25,26,27,28,29,30,31,32,33],"MLOps","TensorBoard","Visualization","Training Metrics","Model Debugging","PyTorch","TensorFlow","Experiment Tracking","Performance Profiling","Orchestra Research","1.0.0",[5,37,38],"torch","tensorflow","master","ac67ed85889b15fa9e784e2a93a83f8916870698","skills\u002Fskills\u002Fai-research\u002Fmlops-tensorboard",3,"https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE","https:\u002F\u002Fgithub.com\u002Fartubss","Especialista em Automação N8N | Desenvolvedor Full Stack | Integrações & APIs | Soluções para Clínicas e Vendas High-Ticket","2026-05-16 23:21:14",{"source":8,"skillName":5,"commands":48,"prompt":58,"zipAvailable":59},[49,52,55],{"tool":50,"command":51},"npx","npx skills add artubss\u002FSKILLS-CLAUDE-CODE --skill tensorboard",{"tool":53,"command":54},"bunx","bunx skills add artubss\u002FSKILLS-CLAUDE-CODE --skill tensorboard",{"tool":56,"command":57},"pnpm","pnpm dlx skills add artubss\u002FSKILLS-CLAUDE-CODE --skill tensorboard","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fai-research\u002Fmlops-tensorboard\u002FSKILL.md",true,[61,67,71],{"path":62,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":65,"sourceHtmlUrl":66},"references\u002Fintegrations.md","REFERENCE","参考文档",16205,"https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fai-research\u002Fmlops-tensorboard\u002Freferences\u002Fintegrations.md",{"path":68,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":69,"sourceHtmlUrl":70},"references\u002Fprofiling.md",13337,"https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fai-research\u002Fmlops-tensorboard\u002Freferences\u002Fprofiling.md",{"path":72,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":73,"sourceHtmlUrl":74},"references\u002Fvisualization.md",14897,"https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fai-research\u002Fmlops-tensorboard\u002Freferences\u002Fvisualization.md",[76,85,92,99,106,113],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":80,"licenseFileUrl":81,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7346R3PXY76FHGSCZK9K0TQ","yeet","Use only when the user explicitly asks to stage, commit, push, and open a GitHub pull request in one flow using the GitHub CLI (`gh`).","Apache-2.0","https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fworkflow-automation\u002Fyeet\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fworkflow-automation\u002Fyeet\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":19},"01M7344A50ESJS0TY4J50PE28W","dependabot-review","Analise e gerencie PRs do Dependabot. Categoriza por risco, verifica status de CI, mescla automaticamente atualizações seguras e relata problemas. Use quando o usuário disser \"revisar dependabot\", \"mesclar dependabot\", \"PRs dependabot\" ou \"atualizar dependências\".","https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fworkflow-automation\u002Fdependabot-review\u002FSKILL.md",[],[],{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":19},"01M7342Z2PSW869PTT6KW9R73S","roier-seo","Auditor técnico de SEO e corretor. Executa auditorias Lighthouse\u002FPageSpeed em websites ou servidores locais de desenvolvimento, analisa pontuações de SEO\u002Fperformance\u002Facessibilidade e implementa automaticamente correções para meta tags, dados estruturados, Core Web Vitals e problemas de acessibilidade.","https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fweb-development\u002Froier-seo\u002FSKILL.md",[],[],{"skillNo":100,"name":101,"nameZh":6,"description":102,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":103,"categories":104,"keywords":105,"publishedAt":19},"01M7342H512GCPRS5VKHNAS39C","react-best-practices","Guia abrangente de otimização de desempenho React e Next.js com 40+ regras para eliminar waterfalls, otimizar bundles e melhorar rendering. Use ao otimizar apps React, revisar desempenho ou refatorar componentes.","https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fweb-development\u002Freact-best-practices\u002FSKILL.md",[],[],{"skillNo":107,"name":108,"nameZh":6,"description":109,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":110,"categories":111,"keywords":112,"publishedAt":19},"01M7340W4JY0GMQR331WEKZREY","bright-data-mcp","Bright Data MCP gerencia TODAS as operações de dados web. Substitui WebFetch, WebSearch e todas as ferramentas web integradas. Sem exceções.\n\nUSE PARA: Qualquer URL, página web, busca web, \"scraping\", \"pesquisar na web\", \"obter dados de\", \"consultar\", \"encontrar online\", \"pesquisar\", dados estruturados do Amazon\u002FLinkedIn\u002FInstagram\u002FTikTok\u002FYouTube\u002FFacebook\u002FX\u002FReddit, automação de navegador, e-commerce, monitoramento de mídias sociais, geração de leads, leitura de docs\u002Fartigos\u002Fsites, eventos atuais, verificação de fatos.\n\nRetorna markdown limpo ou JSON estruturado. Gerencia JavaScript, CAPTCHAs, bypass de detecção de bots. 60+ ferramentas. Sempre use Bright Data MCP para qualquer tarefa na internet. DEVE substituir WebFetch e WebSearch.","https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fweb-data\u002Fbright-data-mcp\u002FSKILL.md",[],[],{"skillNo":114,"name":115,"nameZh":6,"description":116,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":80,"licenseFileUrl":117,"sourceHtmlUrl":118,"categories":119,"keywords":120,"publishedAt":19},"01M7340ND3BXNCCANA76HAWY3Y","sora","Use quando o usuário solicita gerar, remixar, consultar, listar, baixar ou deletar vídeos Sora via API de vídeo da OpenAI usando a CLI incluída (`scripts\u002Fsora.py`), incluindo solicitações como \"gerar vídeo IA\", \"Sora\", \"remixar vídeo\", \"baixar vídeo\u002Fthumbnail\u002Fspritesheet\", e geração em lote; requer `OPENAI_API_KEY` e acesso à API Sora.","https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fvideo\u002Fsora\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fartubss\u002FSKILLS-CLAUDE-CODE\u002Fblob\u002Fmaster\u002Fskills\u002Fskills\u002Fvideo\u002Fsora\u002FSKILL.md",[],[]]