File size: 2,563 Bytes
6866fdc | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 | ---
license: apache-2.0
library_name: coreml
tags:
- coreml
- audio
- speech-enhancement
- noise-suppression
- streaming
- dpdfnet
---
# DPDFNet-4 Core ML
Stateful Core ML and MLX assets for **DPDFNet-4**, CEVA's 16 kHz real-time speech-enhancement model. This repository is the default DPDFNet-4 source for the [`DeepFilterNetCoreML`](https://github.com/kylehowells/DeepFilterNet-mlx) Swift product.
## Origin
- Official implementation: [ceva-ip/DPDFNet](https://github.com/ceva-ip/DPDFNet)
- Official models: [Ceva-IP/DPDFNet](https://huggingface.co/Ceva-IP/DPDFNet)
- Paper: [DPDFNet](https://arxiv.org/abs/2512.16420)
- Swift runtime and conversions: [kylehowells/DeepFilterNet-mlx](https://github.com/kylehowells/DeepFilterNet-mlx)
- Conversion script: [`Scripts/Conversion/convert_dpdfnet_to_coreml.py`](https://github.com/kylehowells/DeepFilterNet-mlx/blob/feature/deepfilternet4/Scripts/Conversion/convert_dpdfnet_to_coreml.py)
## Runtime contract
| Property | Value |
|---|---:|
| Sample rate | 16,000 Hz |
| Input hop | 160 samples / 10 ms |
| FFT | 320 samples |
| DPRNN blocks | 4 |
| Fixed algorithmic delay | 800 samples / 50 ms |
## Recommended model
`DPDFNet4-Streaming-FP32.mlpackage` is the production default. Its conversion validation against the PyTorch graph produced 0.9999999999 output correlation. The FP16 explicit-state and resident-state variants are included for device-specific experiments but are not selected automatically because their fidelity is lower.
Current 60-second Swift/Core ML validation: 13.953 seconds total (4.30x real time), with 0.99999999 waveform correlation to official ONNX. Live p50/p95 model processing latency was 3.686/4.171 ms on the development Apple Silicon Mac. Device results vary.
## Swift usage
```swift
import DeepFilterNetCoreML
let enhancer = try await DeepFilterNetCoreMLStreamer.load(
configuration: .init(variant: .dpdfNet4)
)
let output = try enhancer.processHop(input160Samples)
```
## Files
- `DPDFNet4-Streaming-FP32.mlpackage`: recommended explicit FP32 state graph.
- `DPDFNet4-Streaming.mlpackage`: explicit FP16 state graph.
- `DPDFNet4-Streaming-State.mlpackage`: Core ML `MLState` graph for supported OS versions.
- `DPDFNet4-initial-state-f32.bin`: canonical recurrent-state initialization.
- `config.json` and `model.safetensors`: matching Swift MLX/DSP model assets.
- `conversion-report.json`: graph-level validation and measured conversion latency.
## License
Apache-2.0, matching the official CEVA DPDFNet repository. See `LICENSE` and the original project.
|