VALDI

Ready to contribute to cutting-edge AI while earning rewards? This guide helps Flexstack users seamlessly deploy a Flexstack AI Host on Valdi.ai, a powerful cloud platform ideal for AI and machine learning.

Before You Start:

• VALDI Account: Make sure you have a Valdi.ai account with sufficient credits. Sign up and top up your account if needed (pricing details available here)

• Flexstack AI: Sign up for a free Flexstack Network account at https://flexstack.ai to obtain your username and password.

Step 1: Configure Your Valdi.ai Environment

1. GPU Cloud Creation: Head to the valdi.ai console and choose "Available Servers" in "Cloud GPUs".

2. Select Resources: We recommend 1x RTX A4000 16GB VRAM for image generation tasks. 1x GeForce RTX 4090 24GB VRAM for deploying LLMs.

The Cloud GPUs available on valdi.ai

1. Configure Server: Proceed with configuring your server with details such as Host Name, number of GPUs, number of CPU cores, and RAM (recommend 4 cores 16GB RAM) and OS (Recommendation: Choose the pre-setup OS as Ubuntu 20.04 + PyTorch).

Configure your virtual machine.

"Ubuntu 20.04 + PyTorch" OS is recommended.

1. Starting your server: Review and click "Launch virtual machine" to begin. (NOTE: SSH access information will be displayed in a popup. You should remember and save it).

Your server SSH login information.

Step 2: Deploy Your Flexstack AI Host

1. Clone and Navigate: Connect to your server via SSH using the terminal.

Run the following command to clone the Flexstack AI Host code:

git clone https://github.com/parrotnetwork/parrot-host

1. Move and Install: Navigate to the Flexstack AI Host directory:

cd parrot-host

Then, install required packages:

pip install -r requirements.txt

1. Authentication and Environment: Obtain your Flexstack Network credentials from https://flexstack.ai/ and run the following command, replacing with your actual username and password:

python get_token_and_create_env.py --username <your_username> --password <your_password>

1. Verification: Ensure the .env file is created (check with ls -a). This file stores your credentials securely.

2. Choose Your Task: Select the Flexstack AI task you wish to deploy

• For Image Generation task (with Stable Diffusion 1.5, SDXL, or SDXL Lightning model) run:

sh scripts/parrot_{sd/sdxl/sdxl_lightning}.sh

For Video Generation task, run:

sh scripts/parrot_txt2vid.sh

For LLMs task (with google/gemma-7b model), run:

sh scripts/parrot_llm_gemma.sh

For SD LoRA training task, run:

sh scripts/parrot_lora.sh

You're In! Congratulations! Your Flexstack Host is now running on Valdi.ai, actively contributing to the Flexstack AI network and potentially earning rewards based on your usage.

Additional Tips:

• Refer to the official Flexstack AI Host documentation for further details and troubleshooting.

• Join the Flexstack Network community forum for discussions, support, and updates.

By deploying your Flexstack AI Host on Valdi.ai, you're joining a vibrant community pushing the boundaries of AI. Welcome aboard!