Text2Audio / requirements.txt
IFMedTechdemo's picture
Update requirements.txt
d19b753 verified
raw
history blame
1.25 kB
# Core dependencies (pinned for ZeroGPU; Oct 2025 compatible)
torch==2.5.1
torchaudio==2.5.1
torchvision==0.20.1 # Matches torch 2.5.1
numpy==2.0.0 # Stable 2.x; works with PyTorch 2.5+
librosa==0.11.0 # Supports NumPy 2.x
soundfile==0.13.1
scipy>=1.11.0
pandas>=2.0.0
requests>=2.28.0
tqdm>=4.65.0
# HF Ecosystem (compatible with PyTorch 2.5.1)
transformers>=4.44.0 # Flexible; supports 2.5.1
diffusers>=0.30.0 # Avoids pre-2.5 bugs
accelerate>=0.33.0 # Latest stable
fsspec==2024.9.0 # Fixes common conflicts
safetensors>=0.4.0
datasets>=2.14.0
# NeuTTS-Air specific (loosened for 2.5.1)
neucodec>=0.0.3
phonemizer==3.3.0
vector-quantize-pytorch==1.17.8
torchao>=0.3.0 # Updated for 2.5
torchtune>=0.3.0
resemble-perth==1.0.1
# Kokoro TTS
kokoro>=0.2
bitsandbytes>=0.41.0 # Stable with 2.5.1; avoid 0.43+ if errors
snac>=0.1.0
pyttsx3>=3.1
gitpython>=3.1.0
# Chatterbox Multilingual (merged; 2.5.1 compatible)
resampy==0.4.3
s3tokenizer>=0.1.0
omegaconf==2.3.0
silero-vad==6.0.0 # NumPy 2.x compatible via torch
conformer==0.3.2
# Optionals (uncomment for languages)
# spacy_pkuseg>=0.0.29 # Chinese
# pykakasi>=2.2.0 # Japanese
# russian-text-stresser @ git+https://github.com/Vuizur/add-stress-to-epub
# dicta-onnx>=0.1.0 # Hebrew