> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sulie.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Installation

This guide explains how to install the Sulie SDK and its dependencies. The SDK
provides a Python interface to interact with Sulie's forecasting capabilities.

<Info>
  **Prerequisite** You should have installed Python (version 3.8 or higher).
</Info>

## SDK Installation

You can install the Sulie SDK using pip:

```bash theme={null}
pip3 install sulie
```

For a specific version of the SDK, specify the version number:

```bash theme={null}
pip3 install sulie==1.0.8
```

## Dependencies

The Sulie SDK requires the following Python packages, which are automatically
installed when you install the SDK:

* `requests`: For making HTTP requests to the Sulie API.
* `pandas`: For data manipulation and handling time series data.
* `pyarrow`: For efficient data serialization and transfer.
* `tqdm`: For progress tracking during data upload and processing.

## Verifying Installation

You can verify your installation by importing the SDK in Python:

```python theme={null}
import sulie
print(sulie.__version__)
```

This should print the version number of your installed SDK.

## Troubleshooting

Here's how to solve some common problems when working with the SDK.

<AccordionGroup>
  <Accordion title="Dataset does not exist">
    Make sure you create a dataset first using the SDK:

    ```python theme={null}
    import pandas as pd
    from sulie import Sulie

    client = Sulie(
        api_key=os.environ.get("SULIE_API_KEY")
    )

    # Read the dataset
    df = pd.read_csv(...)

    dataset = client.upload_dataset(
        df=df,
        name="wind-power-generation-mw",
        mode="append"  # Push mode, `append` or `overwrite`
    )
    ```
  </Accordion>

  <Accordion title="Model not found">
    When fine-tuning the Mimosa foundation model, it is important to wait for
    the confirmation email sent to administrators of your Sulie organization
    acknowledging that fine-tuning has completed. This may take up to an hour.
  </Accordion>
</AccordionGroup>