-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathrequirements.txt
More file actions
40 lines (35 loc) · 885 Bytes
/
Copy pathrequirements.txt
File metadata and controls
40 lines (35 loc) · 885 Bytes
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
# Training-only dependencies for openWakeWord custom model training.
# Install inside WSL2 / Linux:
# pip install -r requirements.txt
# --- Core training ---
openwakeword
torch>=2.0
torchaudio>=2.0
torchinfo>=1.8
torchmetrics>=1.2
speechbrain>=0.5.14
# --- Audio augmentation ---
audiomentations>=0.33
torch-audiomentations>=0.11
acoustics>=0.2.6
# --- TTS & phonemics ---
piper-phonemize>=1.1.0
pronouncing>=0.2.0
deep-phonemizer>=0.0.19
# --- Data & utilities ---
datasets>=2.14
mutagen>=1.47
webrtcvad>=2.0.10
scipy
numpy
pyyaml
tqdm
requests
# --- Compat patches ---
# setuptools<82 keeps pkg_resources (removed in 82+)
# soundfile is the fallback backend for patched torchaudio 2.10+
setuptools<82
soundfile
onnxruntime # for verify-model step
onnx # for ONNX export from PyTorch
onnxscript # required by torch.onnx.export in PyTorch 2.x