voiceagents-cosyvoice2 / environment-comprehensive.yml
starkprince's picture
Upload folder using huggingface_hub
778d4b8 verified
name: gsva-python310
channels:
- defaults
- conda-forge
- pytorch
dependencies:
# Python version
- python=3.10
# Core dependencies from conda
- pip
- numpy=1.26.4
- setuptools=69.5.1
- cython
- packaging
- wheel
# CUDA toolkit for GPU support
- cudatoolkit=12.1
# Pip dependencies
- pip:
# Core ML frameworks
- torch==2.3.1
- torchaudio==2.3.1
- torchvision
- transformers==4.43.4
- tokenizers==0.19.1
# LLaMA-Omni2 specific
- sentencepiece==0.1.99
- shortuuid==1.0.13
- accelerate==0.33.0
- peft==0.14.0
- bitsandbytes==0.45.0
- flash-attn
# Web frameworks
- gradio==5.3.0
- gradio_client==1.4.2
- fastapi==0.115.11
- uvicorn==0.30.0
- httpx==0.28.1
- requests==2.32.3
# Audio processing
- openai-whisper==20231117
- conformer==0.3.2
- librosa==0.10.2
- matcha-tts
# CosyVoice dependencies
- HyperPyYAML==1.2.2
- onnxruntime-gpu==1.18.0
- onnx==1.16.0
- ruamel.yaml
# ML utilities
- einops==0.6.1
- einops-exts==0.0.4
- timm==0.6.13
- scikit-learn==1.2.2
- datasets==2.18.0
- diffusers==0.27.2
- lightning==2.2.4
- omegaconf==2.3.0
- hydra-core==1.3.2
# Utilities
- huggingface-hub==0.25.1
- hf_transfer
- ninja==1.11.1.3
- pydantic==2.7.0
- markdown2[all]==2.5.2
- gdown==5.1.0
- wget==3.2
- inflect==7.3.1
- WeTextProcessing==1.0.3
- matplotlib==3.10.1
# Install from GitHub
- git+https://github.com/FunAudioLLM/CosyVoice.git
# Local package (editable install)
- -e .
variables:
# CUDA environment variables
PATH: /usr/local/cuda/bin:$PATH
LD_LIBRARY_PATH: /usr/local/cuda/lib64:$LD_LIBRARY_PATH
# Hugging Face configuration
HF_HUB_ENABLE_HF_TRANSFER: 1
HF_HOME: ~/.cache/huggingface
# PyTorch configuration
TORCH_CUDA_ARCH_LIST: "7.0;7.5;8.0;8.6;8.9;9.0"
MAX_JOBS: 4