> For the complete documentation index, see [llms.txt](https://docs.cloudeka.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.cloudeka.ai/guidance-for-enterprise/deka-gpu-mlops/pipelines/create-pipelines.md).

# Create Pipelines

First, click the **+ Upload Pipeline** button in the top-right corner of the page to start creating a pipeline.

<figure><img src="/files/IV5rZ8F9miJG3jZohiqf" alt=""><figcaption></figcaption></figure>

You will be directed to the **New Pipeline** window, where you need to complete several sections before clicking the **Create** button.

<figure><img src="/files/KiLcw2H53rp93OyceWdz" alt=""><figcaption></figcaption></figure>

a. **Upload pipeline or pipeline version**

Choose whether to create a new pipeline or create a new version under an existing pipeline. If you select **Create a new pipeline**, you will be prompted to enter the pipeline name and description. If you select **Create a new pipeline version under an existing pipeline**, the existing pipeline will be updated with the newly imported version. In this case, you must select the pipeline that will be updated. In both cases, the pipeline can be imported by uploading a **.yaml**, **.zip**, or **.tar.gz** file, or by providing a URL that contains one of these file formats.

b. **Select if the new pipeline will be private or shared**

Choose either **Private** or **Shared**. Private pipelines appear in the **Private** section of the Pipelines menu, while shared pipelines appear in the **Shared** section.

c. **Upload pipeline with the specified package**

Enter the **Pipeline Name**. The **Description** field is optional. Then, choose either **Upload a File** or **Import by URL**. After completing the required configuration, click the **Create** button.

<figure><img src="/files/IoYpKVDzeatzNwgqzcyX" alt=""><figcaption></figcaption></figure>

Once the pipeline is created, you will be redirected to the **Pipelines** page, where the available pipelines are listed. Click the name of the pipeline that you just created to open its details page.

<figure><img src="/files/mYFvay2yjJTkCrnjkpKv" alt=""><figcaption></figcaption></figure>

A graph visualization of the pipeline is displayed. This visualization is similar to a pipeline created using the Python SDK. Initially, a summary of the imported pipeline is displayed. If you close the summary, a detailed graph view of the pipeline is shown. To display the summary again, click the **Show Summary** button in the bottom-left corner.

<figure><img src="/files/AvYljqIK4SAgIHGPHUuP" alt=""><figcaption></figcaption></figure>

You can download the imported pipeline as a YAML file by clicking the **YAML** button. In the top-right corner, three additional options are available:

<figure><img src="/files/ZmD7LmOMnsMU408iiE02" alt=""><figcaption></figcaption></figure>

1. [**Create Run**](/guidance-for-enterprise/deka-gpu-mlops/runs/create-run.md)\
   This option opens a form for creating a pipeline run. You can configure and execute the pipeline workflow by providing the required information and parameters.
2. **Upload version**\
   This option opens a form for uploading a new version of the pipeline. The interface is the same as the one used when selecting **Create a new pipeline version under an existing pipeline** during pipeline import.
3. [**Create Experiment**](/guidance-for-enterprise/deka-gpu-mlops/experiments-kfp/create-experiment.md)\
   This option opens a form for creating an experiment. The interface is similar to the one used when creating an experiment from the **Experiments (KFP)** menu. After entering the experiment name and description, you can configure the pipeline run settings in the same way as when creating a new run.