Model Summary
MLP model trained on Ames Housing Dataset in ONNX format
Usage
See the included notebook for details.
System
This is a standalone ONNX model supporting inference using frameworks such as ONNX Runtime.
Implementation requirements
Trained with MindSpore 2.8.0 + CANN 8.5.0 on OrangePi AIpro (20T) featuring a single Ascend 310B1 NPU chip and core
Model Characteristics
Model initialization
Trained from scratch
Model stats
MLP with 1 hidden layer and dropout.
- Hidden layer: 288 input channels, 64 output channels, ReLU activation
- Dropout layer with
p=0.2
- Fully connected layer: 64 input channels, 1 output channel
