Organiza tus páginas con colecciones
Guarda y categoriza el contenido según tus preferencias.
Puedes usar el SDK de Vertex AI para Python o la consola de Google Cloud para crear o borrar un experimento. El SDK es una biblioteca de código de Python que puedes usar para crear y administrar experimentos de manera programática. La consola es una interfaz de usuario basada en la Web que permite crear y administrar experimentos de forma visual.
Crea el experimento con una instancia de TensorBoard
SDK de Vertex AI para Python
Crea un experimento y asocia, opcionalmente, una instancia de TensorBoard de Vertex AI con el SDK de Vertex AI para Python. Agrega una descripción al experimento para documentar su propósito. Consulta init en la documentación de referencia del SDK de Vertex AI.
experiment_name: Proporciona un nombre para tu experimento.
experiment_description: Proporciona una descripción para tu experimento.
experiment_tensorboard: Opcional. La instancia de Vertex TensorBoard que se usa como TensorBoard de respaldo para el experimento proporcionado.
Si no se proporciona un experiment_tensorboard, se genera una instancia de TB predeterminada y usa en este experimento. Nota: Si las CMEK (claves de encriptación) necesitan estar asociadas con la instancia de TensorBoard, experiment_tensorboard ya no es opcional.
project: . Puedes encontrar estos IDs en la página de bienvenida de la consola de Google Cloud .
location: Consulta la Lista de ubicaciones disponibles. Asegúrate de usar una región compatible con TensorBoard si creas una instancia de TensorBoard.
Google Cloud console
Usa estas instrucciones para crear un experimento.
En la consola de Google Cloud , ve a la página Experiments. Ir a Experimentos
Asegúrate de estar en el proyecto en el que deseas crear el experimento.
Haz clic en add_box Crear para abrir el panel Experimento. Aparecerá el panel Crear experimento.
En el campo Nombre del experimento, proporciona un nombre para identificar de forma única tu experimento.
Opcional. En el campo Instancia de TensorBoard, selecciona una instancia del menú desplegable o proporciona un nombre para la instancia nueva de TensorBoard.
Haz clic en Crear para crear el experimento.
Crea un experimento sin una instancia predeterminada de TensorBoard
SDK de Vertex AI para Python
Crea un experimento Agrega una descripción al experimento para documentar su propósito. Consulta init en la documentación de referencia del SDK de Vertex AI.
experiment_name: Proporciona un nombre para tu experimento.
experiment_description: Proporciona una descripción para tu experimento.
project: . Puedes encontrar estos IDs en la página de bienvenida de la consola de Google Cloud .
location: Consulta la Lista de ubicaciones disponibles. Asegúrate de usar una región compatible con TensorBoard si creas una instancia de TensorBoard.
Borrar experimento
Borrar un experimento borra ese experimento y todos los que se ejecutan con él. El experimento de Vertex AI TensorBoard asociado con el experimento no se borra. Para borrar un experimento de TensorBoard, consulta Borra el experimento de Vertex AI TensorBoard desactualizado.
Además, no se quitan las ejecuciones de canalizaciones, los artefactos y ni las ejecuciones asociadas con el experimento borrado. Puedes encontrarlos en la consola de Google Cloud .
En el caso de los artefactos y las ejecuciones, el servicio de Vertex ML Metadata estipula un cargo mensual de USD 10 por GB.
delete_backing_tensorboard_runs: Si es verdadero, también se borrarán las ejecuciones de TensorBoard de Vertex AI asociadas con el experimento que se ejecuta en este experimento que usamos para almacenar las métricas de series temporales.
Console
Usa las siguientes instrucciones para borrar un experimento.
En la consola de Google Cloud , ve a la página Experiments. Ir a Experimentos
Selecciona la casilla de verificación asociada con el experimento que deseas borrar. Aparecerá la opción Borrar.
Haz clic en Borrar.
Como alternativa, puedes ir al menú de opciones more_vert que se encuentra en la misma fila que el experimento y seleccionar borrar.
Visualiza la lista de experimentos en la consola de Google Cloud
En la Google Cloud consola, en la sección Vertex AI, ve a la página Experimentos.
En la vista Seguimiento de experimentos, aparece una lista de los experimentos de tu proyecto. Si asociaste una instancia de TensorBoard de Vertex AI con tu experimento, aparecerá en la lista como Backing Tensorboard Experiment “tu-experimento ”.
[[["Fácil de comprender","easyToUnderstand","thumb-up"],["Resolvió mi problema","solvedMyProblem","thumb-up"],["Otro","otherUp","thumb-up"]],[["Difícil de entender","hardToUnderstand","thumb-down"],["Información o código de muestra incorrectos","incorrectInformationOrSampleCode","thumb-down"],["Faltan la información o los ejemplos que necesito","missingTheInformationSamplesINeed","thumb-down"],["Problema de traducción","translationIssue","thumb-down"],["Otro","otherDown","thumb-down"]],["Última actualización: 2025-09-04 (UTC)"],[],[],null,["# Create or delete an experiment\n\nYou can use either the Vertex AI SDK for Python or the Google Cloud console to\ncreate or delete an experiment. The SDK is a library of Python code that you\ncan use to programmatically create and manage experiments. The console is a\nweb-based user interface that you can use to create and manage experiments\nvisually.\n| When creating an experiment using the Google Cloud console for the first time, be sure that there's a `default` Metadata Store. To check, go to your project's **Metadata** page in the Google Cloud console. See [Configure your project's metadata store](/vertex-ai/docs/ml-metadata/configure)\n\nCreate experiment with a TensorBoard instance\n---------------------------------------------\n\n### Vertex AI SDK for Python\n\n\nCreate an experiment and, optionally, associate a Vertex AI TensorBoard instance using\nthe Vertex AI SDK for Python. Add a description for the\nexperiment to document its purpose. See [`init`](/python/docs/reference/aiplatform/latest/google.cloud.aiplatform#google_cloud_aiplatform_init)\nin the Vertex AI SDK reference documentation. \n\n### Python\n\n from typing import Optional, Union\n\n from google.cloud import aiplatform\n\n\n def create_experiment_sample(\n experiment_name: str,\n experiment_description: str,\n experiment_tensorboard: Optional[Union[str, aiplatform.Tensorboard]],\n project: str,\n location: str,\n ):\n aiplatform.init(\n experiment=experiment_name,\n experiment_description=experiment_description,\n experiment_tensorboard=experiment_tensorboard,\n project=project,\n location=location,\n )\n\n- `experiment_name`: Provide a name for your experiment.\n- `experiment_description`: Provide a description for your experiment.\n- `experiment_tensorboard`: Optional. The Vertex TensorBoard instance to use as a backing TensorBoard for the provided experiment. If no `experiment_tensorboard` is provided, a default TB instance is created and used by this experiment. Note: If CMEK (encryption keys) need to be associated with the TensorBoard instance, then `experiment_tensorboard` is no longer optional.\n- `project`: . You can find these IDs in the Google Cloud console [welcome](https://console.cloud.google.com/welcome) page. \n- `location`: See [List of available locations](/vertex-ai/docs/general/locations) Be sure to use a region that supports TensorBoard if creating a TensorBoard instance.\n\n### Google Cloud console\n\n\nUse these instructions to create an experiment.\n\n1. In the Google Cloud console, go to the **Experiments** page. \n [Go to Experiments](https://console.cloud.google.com/vertex-ai/experiments)\n2. Be sure you're in the project you want to create the experiment in. \n3. Click **add_box\n Create** to open the **Experiment** pane. The **Create experiment** pane appears.\n4. In the **Experiment name** field, provide a name to uniquely identify your experiment.\n5. Optional. In the **TensorBoard instance** field, select an instance from the drop-down or provide a name for your new TensorBoard instance.\n6. Click **Create** to create your experiment.\n\nCreate an experiment without a default TensorBoard instance\n-----------------------------------------------------------\n\n### Vertex AI SDK for Python\n\n\nCreate an experiment. Add a description for the\nexperiment to document its purpose. See [`init`](/python/docs/reference/aiplatform/latest/google.cloud.aiplatform#google_cloud_aiplatform_init)\nin the Vertex AI SDK reference documentation. \n\n### Python\n\n from google.cloud import aiplatform\n\n\n def create_experiment_without_default_tensorboard_sample(\n experiment_name: str,\n experiment_description: str,\n project: str,\n location: str,\n ):\n aiplatform.init(\n experiment=experiment_name,\n experiment_description=experiment_description,\n experiment_tensorboard=False,\n project=project,\n location=location,\n )\n\n- `experiment_name`: Provide a name for your experiment.\n- `experiment_description`: Provide a description for your experiment.\n- `project`: . You can find these IDs in the Google Cloud console [welcome](https://console.cloud.google.com/welcome) page. \n- `location`: See [List of available locations](/vertex-ai/docs/general/locations) Be sure to use a region that supports TensorBoard if creating a TensorBoard instance.\n\nDelete experiment\n-----------------\n\nDeleting an experiment deletes that experiment and all experiment runs\nassociated with the experiment. The Vertex AI TensorBoard experiment\nassociated with the experiment is not deleted. To delete a TensorBoard\nexperiment, see\n[Delete outdated Vertex AI TensorBoard experiment](/vertex-ai/docs/experiments/user-journey/uj-delete-outdated-tb-experiments).\n\nAlso, any pipeline runs, artifacts, and executions associated with the deleted\nexperiment are not removed. These can be found in the Google Cloud console.\nFor artifacts and executions, a $10/GB monthly charge is handled by the\nVertex ML Metadata service. \n\n### Vertex AI SDK for Python\n\nThe following sample uses the\n[`delete`](/python/docs/reference/aiplatform/latest/google.cloud.aiplatform.ExperimentRun#google_cloud_aiplatform_ExperimentRun_delete)\nmethod from the\n[`ExperimentClass`](/python/docs/reference/aiplatform/latest/google.cloud.aiplatform.ExperimentRun).\n\n### Python\n\n from google.cloud import aiplatform\n\n\n def delete_experiment_sample(\n experiment_name: str,\n project: str,\n location: str,\n delete_backing_tensorboard_runs: bool = False,\n ):\n experiment = aiplatform.Experiment(\n experiment_name=experiment_name, project=project, location=location\n )\n\n experiment.delete(delete_backing_tensorboard_runs=delete_backing_tensorboard_runs)\n\n- `experiment_name`: Provide a name for your experiment.\n- `project`: . You can find these IDs in the Google Cloud console [welcome](https://console.cloud.google.com/welcome) page.\n- `location`: See [List of available locations](/vertex-ai/docs/general/locations)\n- `delete_backing_tensorboard_runs`: If True will also delete the Vertex AI TensorBoard runs associated with the experiment runs under this experiment that we used to store time series metrics.\n\n### Console\n\n\nUse the following instructions to delete an experiment.\n\n1. In the Google Cloud console, go to the **Experiments** page. \n [Go to Experiments](https://console.cloud.google.com/vertex-ai/experiments)\n2. Select the checkbox associated with the experiment you want to delete. The **Delete** option appears.\n3. Click **Delete** .\n - Alternatively, you can go to the more_vert options menu that is in the same row as the experiment and select **delete**.\n\nView list of experiments in Google Cloud console\n------------------------------------------------\n\n1. In the Google Cloud console, in the Vertex AI section, go to the\n **Experiments** page.\n\n [Go to the Experiments page](https://console.cloud.google.com/vertex-ai/experiments)\n2. Check to be sure you are in the correct project.\n\n3. A list of experiments for your project appears in\n the **Experiment tracking** view. \n\n If you associated a Vertex AI TensorBoard instance with your\n experiment it shows up in the list as \"*your-experiment* Backing\n TensorBoard Experiment\".\n\nWhat's next\n-----------\n\n- [Create and manage experiment runs](/vertex-ai/docs/experiments/create-manage-exp-run)\n- [Delete outdated Vertex AI TensorBoard experiment](/vertex-ai/docs/experiments/user-journey/uj-delete-outdated-tb-experiments)\n\n### Relevant notebook sample\n\n- [Model training with prebuilt data pre-processing code](/vertex-ai/docs/experiments/user-journey/uj-model-training)"]]