Deploy and develop AI model with the FastAPI

I have a text classification model that uses Simpletransformer to train. I want to serve it as a web service with the FastAPI and dockersie it. Deploy it on an AWS cloud instance and make sure the model is served with GPU power.

A few things to do

1. build the RestAPI for the model
the model will take two parameters, a text plus an array, a probability score for the return
example request:
{
"text1": "texta",
"text2": ["texta","textb","textc"]
}
and example response
{
"ai_classifier_score": [
{
"text1": "texta",
"classifier_score": 0.9946354627609253
},
{
"text1": "textb"",
"classifier_score": 0.0025506913661956787
},
{
"text1": "textc",
"classifier_score": 0.0025454163551330566
}
]
}
2. Dockerise the project with best practices and standards(well-organised file and naming, proper comments)
3. Deploy it on the AWS cloud and ensure the model serves with GPU power.

material provided
model zip file
AWS instance and account

Ref.
https://simpletransformers.ai/docs/installation/
https://fastapi.tiangolo.com/