Cache layer with docker to speed up the development workflow

### Describe your idea/feature/enhancement
[Project: `sam init - python 3.8 - image container `(not zip file)]
`sam build` always rebuild the image even if:
- `requirements.txt` is not changed
- the code is not changed

### Proposal
At least in development, we could change to Dockerfile to cache the layers that didn't changed, such as requirments.txt. When someone has a slow connection, it take A LOT of time.

In order to do this, the default Dockerfile should be changed in:
**New** `Dockerfile`
```
FROM public.ecr.aws/lambda/python:3.8 AS build

COPY requirements.txt ./
RUN python3.8 -m pip install -r requirements.txt -t .

COPY app.py ./
COPY model /opt/ml/model

CMD ["app.lambda_handler"]
```

**Old** `Dockerfile`
```
FROM public.ecr.aws/lambda/python:3.8

COPY app.py requirements.txt ./
COPY model /opt/ml/model

RUN python3.8 -m pip install -r requirements.txt -t .

CMD ["app.lambda_handler"]
```

### Additional Details

Start by locating the default Dockerfile used by `sam init` for the Python 3.8 image-container project, then compare its COPY and pip-install layers with the proposed Dockerfile. Check how `sam build` uses this template and identify the relevant template tests, if present. Done means unchanged requirements and source can reuse Docker cache during development without changing the generated application's behavior.

Written by the indexing model from the issue text.