AI Deploy allows you to deploy AI apps or models. To test or use the product, you can build on existing AI models. For example, you can rely on open-source models or apps.
AI Deploy is covered by OVHcloud Public Cloud Special Conditions.
Portfolio of AI apps and models
To test AI Deploy, you can quickly deploy apps based on those proposed in our portfolio.
Quick examples
Hello world
- Description: Launch your first API with Flask
- Documentation: AI Deploy - Getting started
- Dockerfile: Dockerfile - Hello world
-
Docker image:
ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/ai-deploy-hello-world -
CLI command:
ovhai app run ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/ai-deploy-hello-world - Usage: API - interact with the API with a curl command or a Python script
EDA and interactive prediction
- Description: Explore, analyze iris data, and do interactive prediction with Streamlit
- Documentation: AI Deploy - Tutorial - Deploy an interactive app for EDA and prediction using Streamlit
- Dockerfile: Dockerfile - EDA and prediction on iris data
-
Docker image:
ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/streamlit-eda -
CLI command:
ovhai app run ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/streamlit-eda - Usage: Web interface - access to the app with the URL
Sketch recognition
- Description: Recognize handwritten digits with Gradio
- Documentation: AI Deploy - Tutorial - Deploy a Gradio app for sketch recognition
- Dockerfile: Dockerfile - Sketch recognition
-
Docker image:
ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/gradio-sketch-recognition -
CLI command:
ovhai app run ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/gradio-sketch-recognition - Usage: Web interface - access to the app with the URL
Spam classification
- Description: Classify spam messages with FastAPI
- Documentation: AI Deploy - Tutorial - Deploy and call a spam classifier with FastAPI
- Dockerfile: Dockerfile - Spam classifier API
-
Docker image:
ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/fastapi-spam-classification -
CLI command:
ovhai app run ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/fastapi-spam-classification -
Usage: API - interact with the API with
<app-url>/docsor curl command
Sentiment analysis
- Description: Analyze text sentiment with Hugging Face models and Flask
- Documentation: AI Deploy - Tutorial - Deploy an app for sentiment analysis with Hugging Face and Flask
- Dockerfile: Dockerfile - Sentiment analysis Hugging Face app
-
Docker image:
ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/flask-sentiment-analysis -
CLI command:
ovhai app run ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/flask-sentiment-analysis - Usage: Web interface - access to the app with the URL
Speech-to-Text
- Description: Use Speech-to-Text powers on audio and video
- Documentation: AI Deploy - Tutorial - Create and deploy a Speech to Text application using Streamlit
- Dockerfile: Dockerfile - Speech-to-Text Streamlit app
-
Docker image:
ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/streamlit-speech-to-text -
CLI command:
ovhai app run ij8m5ic1.c1.va1.container-registry.ovh.us/ai-deploy-portfolio/streamlit-speech-to-text - Usage: Web interface - access to the app with the URL
If you want to launch these apps from the OVHcloud control panel, fill in the name of the docker image in Step 2 - Application to deploy.
Each of the following apps launches on port 8080. You don't need to enter it in the launch command.
By default, an app is launched with 1 GPU. However, you can customize the resources you wish to use.
Build you own apps and models to deploy
Below are examples of existing models that can be deployed with AI Deploy. However, you are free to deploy any other AI model or app of your choice.
YOLO
YOLO ('You only look once'), is an Object Detection algorithms family.
References:
- YOLOv5 repository
- YOLOv7 repository
- Access and load YOLOv5 trained models from PyTorch Hub
DALL-E mini
DALL-E mini is an AI model that can draw images from any text prompt (Text-to-Image).
References:
- DALL-E mini repository
- Test DALL-E mini on this website
- Access DALL-E mini model from Hugging Face
Stable Diffusion
Stable Diffusion is Text-to-Image model that generates images from text.
References:
- Stable Diffusion repository
- Stable Diffusion website
- Access Stable Diffusion model from Hugging Face
EfficientNet
EfficientNet is a family of Image Classification models. There are eight different EfficientNet models (b0 -> b7)
References:
- Access and load EfficientNet model from PyTorch Hub
ResNet
ResNet are AI models based residual neural network whose aim is to solve Image Classification tasks.
References:
- Access and load ResNet models from PyTorch Hub
MobileNet V2
MobileNet are Computer Vision models designed to be used in mobile applications. They are known for their small size and low latency.
References:
- Access and load ResNet models from PyTorch Hub
GPT-2
GPT-2 is a Text Generation model developed by Open AI.
References:
- Find more information about GPT-2 here
- Access GPT-2 model from Hugging Face
BERT
Famous NLP models based on BERT can also be deployed for Text Classification for example.
References:
- Access BERT-based models from Hugging Face
BART
BART-based models can also be deployed for Zero-Shot Classification tasks.
References:
- Access BART-based models from Hugging Face
Go further
You can also refer to our GitHub repository to find examples of AI apps to deploy.
For more information and tutorials, please see our other AI & Machine Learning support guides or explore the guides for other OVHcloud products and services.
If you need training or technical assistance to implement our solutions, contact your sales representative or click on this link to get a quote and ask our Professional Services experts for a custom analysis of your project.