COLAB_INSTRUCTIONS.md

# Running dashVectorspace on Google Colab

Since the ingestion process can be compute-intensive, running on Google Colab (especially with a GPU) is a great option.

## Steps

3.  **Setup Environment**:
    Open the notebook `notebooks/dashVector_full_benchmark.ipynb`.
    Since it clones the code directly from Hugging Face, you do **not** need to upload any zip files.
    
    Simply **Run All Cells**.
    The notebook will:
    - Clone `https://huggingface.co/spaces/dashVector/dashVectorSpace`
    - Install dependencies
    - Run the 25k Benchmark
    - Prompt to download artifacts

4.  **Download Artifacts**:
    The notebook will save models to the `models/` directory.
    Zip and download this folder to your local machine:
    ```python
    !zip -r models.zip models
    from google.colab import files
    files.download('models.zip')
    ```
    Then, place these files into `dashVectorspace/models/` and re-deploy.