Training Configuration Parameters
Audio Classification & Gesture Recognition
| Parameter | Description |
|---|---|
| Number of Epochs | Total training iterations over the dataset |
| Patience | Epochs to wait before early stopping if validation loss does not improve |
| Learning Rate | Step size for weight updates during optimisation |
| Enable QAT | Enables Quantization Aware Training |
| Activation Bits | Bit width for activation quantisation |
| Weight Bits | Bit width for weight quantisation |
Image Classification
| Parameter | Description | Default |
|---|---|---|
| Epochs | Total complete passes through the dataset | 100 |
| Patience | Epochs without improvement before auto-stop | 4 |
| Batch Size | Images per training step — reduce if out of memory | 16 |