RIME Cloud Installation

The cloud installation consists of two parts, with an optional third part:

The Web Client
- No installation needed — your RI representative will provide your access details.
The Python SDK
- This is a programmatic interface to the managed K8s cluster.
The Python Package (Optional)
- This contains the command-line interface (CLI) and Python libraries used to run tests on your models and data.

Pre-requisites

Before beginning, please ensure you have the following:

The following system requirements:
- at least 4 GB RAM
- at least 2 GB free disk space
- 3.7 ≤ python ≤ 3.9
- one of the following package managers
  - pip
  - conda
The rime_trial bundle ( download here )
If installing the RIME Python Package (Optional):
- A token.txt file (will be provided by your RI representative)

Access the Web Client

Your RIME web client will be made available at a URL with the following form: https://rime.<YOUR_ORG_NAME>.rime.dev/

Admin User Setup

Initially, a team member from your organization will be designated as the admin user for their RIME cluster.

This user will receive an email from Robust Intelligence with steps for setting up their account.

Adding Other Members

Once the admin user’s account is configured, they will be able to grant access to other users via the Workspace Settings.

More detailed instructions can be found in the User Management Guide.

Install the Python SDK

In addition to the standard python package, Cloud installations support the RIME Python SDK for programmatic integration with your existing ML pipeline!

Installation is simple:

pip install rime-sdk

Create an API Token

An API token is needed to access the Python SDK. These can be created in the web client via the Workspace Settings.

More detailed instructions can be found in the API Authentication Guide.

Test Connection

Run the code snippet below to test your SDK connection:

from rime_sdk import RIMEClient

rime_client = RIMEClient("rime.<YOUR_ORG_NAME>.rime.dev", "<RIME_API_TOKEN>")

If there are no errors, that means you are ready to kickoff some tests!

Install the RIME Python Package (Optional)

Set Up Your Environment

To avoid interfering with your other projects’ dependencies, we will create a new Python virtual environment for RIME.

Download ( link here ) and extract (tar -xzf rime-trial-0.16.14.tar.gz) the rime_trial bundle.
Place the token.txt file inside of rime_trial/.
Step into rime_trial/ — this will be our working directory throughout the installation.
Select a virtual environment manager (currently either pip or Conda) and follow the appropriate instructions below.

Pip Installation
1. Create a new python virtual environment.
```
python -m venv rime-venv
```
2. Activate the virtual environment.
  
  macOS & Linux
```
source rime-venv/bin/activate
```
  Windows
```
.\rime-venv\Scripts\activate
```
3. Update pip and install required Python packages from trial_requirements.txt.
```
pip install --upgrade pip
pip install -r trial_requirements.txt
```
Conda Installation
1. Create a new conda virtual environment. Replace $PYTHON_VERSION_YOU_WISH_TO_USE with the version of Python that you wish to use (e.g., 3.8).
```
conda create --name rime-venv python=$PYTHON_VERSION_YOU_WISH_TO_USE
```
2. Activate the virtual environment.
```
conda activate rime-venv
```
3. Update pip and install required Python packages from trial_requirements.txt.
```
pip install --upgrade pip
pip install -r trial_requirements.txt
```

Generate Requirements Files and Install

Generate the requirements files.

python rime_helper.py generate-rime-requirements

Register your product license.
```
python rime_helper.py update-license
```
Install the RIME Python package.
```
pip install -r rime_requirements.txt
```

Configure environment variables for use in CLI commands:

macOS & Linux

export RIME_UPLOAD_URL=rime-backend.<YOUR_ORG_NAME>.rime.dev
export RIME_FIREWALL_URL=rime-backend.<YOUR_ORG_NAME>.rime.dev
export RIME_API_KEY=<YOUR_API_KEY>

Windows

$Env:RIME_UPLOAD_URL = 'rime-backend.<YOUR_ORG_NAME>.rime.dev'
$Env:RIME_FIREWALL_URL = 'rime-backend.<YOUR_ORG_NAME>.rime.dev'
$Env:RIME_API_KEY = '<YOUR_API_KEY>' 

Install the RIME NLP Dependencies (Optional)

To use RIME for stress testing NLP models, additional Python dependencies must be installed.

pip install -r nlp_requirements.txt

Install the RIME CV Dependencies (Optional)

To use RIME for stress testing Computer Vision models, additional Python dependencies must be installed.

pip install -r cv_requirements.txt