I use the quantize_static to quantize float32 pytorch model to int8 model. However, it does not reduce the memory usage on CPU. Why?

### Describe the issue

I use the quantize_static to quantize float32 pytorch model to int8 model. However, it does not reduce the memory usage on CPU. Why?
The quantized model size is actually smaller, but the memory usage is almost the same (even slightly bigger) than the float32 model on CPU. #onnxruntime #quantization #quantize_static

### To reproduce

pytorch size : 1.5M float32 model

### Urgency

urgent for help

### Platform

Linux

### OS Version

ubuntu

### ONNX Runtime Installation

Built from Source

### ONNX Runtime Version or Commit ID

1.12

### ONNX Runtime API

Python

### Architecture

X64

### Execution Provider

Default CPU

### Execution Provider Library Version

_No response_