Table 12 Processing time required by our best performing system to process a 1-hour long audio using a CPU and GPU bases setup
From: Multiclass audio segmentation based on recurrent neural networks for broadcast domain data
Feat extraction | Inference | Total time | RTF | |
|---|---|---|---|---|
CPU | 1min 28s | 28s | 1min 56s | 0.032 |
GPU | 2s | 1min 30s | 0.025 |