How can I make tfx use tensorflow-cpu rather than the full tensorflow package?

Question

I am trying to build a lightweight docker image with tfx library installed, as the officially released image is pretty heavy in size (~9 GBs).

Here is the code that I am using for the image now:

FROM python:3.8-slim AS base
RUN pip install --upgrade pip
RUN apt-get update && apt-get install -y gcc g++ && rm -rf /var/lib/apt/lists/*
RUN pip install tfx==1.12.0 --no-cache-dir

The image that is built now is around 2.7 GBs in size. However, I feel like most of these packages are for GPUs and not CPUs as tfx is dependent on the main tensorflow package. Is there a way that I use tensorflow-cpu as its dependency rather than the full package to make the image smaller?

Thanks.