Epeuva cli

Installation (virtualenv)

$ git clone [email protected]:DT42/epeuva-cli.git
$ python3 -m venv env
$ source env/bin/activate
(env)$ cd epeuva-cli
(env)$ python3 setup.py install

---

Prequisites

Set up Epeuva host

To use the Epeuva CLI, you need to have the access to an Epeuva (the End-to-End AI gateway) instance.

Currently, Epeuva has not been open-sourced yet, but it is planned to be (coming soon).

Create config

The Epeuva CLI will load the config file ~/.epeuva-cli.conf.

If this config file is not present or not in the correct format, a prompt will be shown when running any Epeuva CLI command to ask if you want to generate a new config (to replace the old config if any)

$ epeuva
[Errno 2] Config file not found: '/home/user/.epeuva-cli.conf'

Would you like to generate/update the config? [y/N]:

Enter y and the CLI will ask for the Epeuva host url (API url, ex. https://epeuva.ai/api/v1), username and password. When you finished, you will see this line:

Config file created. Please run the CLI command again.

Now you can start using Epeuva CLI.

---

Usage

Basics

Show help

$ epeuva --help

List all Apps

$ epeuva app list

Get single app info

$ epeuva app detail --id 1 show

List all base models

$ epeuva base list

List all labels of a base model

$ epeuva label list --base-id 1

API to upload training data

Upload images to target label

$ epeuva label detail --id 3 upload_images *.jpg

API to trigger training pipeline on Epeuva

Retrain model

$ epeuva job retrain --model-name cli-test-00 \
    --label-list label.txt \
    --app-id 1 \
    --note "Hello world"

API to convert model for a specific hardware

Convert model

$ epeuva job convert --model-name cli-test-00 --model-id 836 --platform snpe --app-id 1

Check task status

Track task status

$ epeuva task detail --uuid <task-id> show

API to download trained model

Download original model

$  epeuva model detail --id 836 --app-id 1 download

Download converted model

$ epeuva model detail --id 836 --app-id 1 download --converted