Installation

Pre-requisites

Install Miniforge or another Conda distribution.

Install DAS

Create an isolated environment. Conda provides Python, Qt, FFmpeg, numcodecs, and uv; uv installs DAS and its remaining Python dependencies. Installing Qt and FFmpeg together through Conda keeps their native libraries compatible.

Users who need the TensorFlow backend should follow the TensorFlow installation instructions for the final TensorFlow-backed release.

On Windows and Linux:

conda create -n das -c conda-forge --strict-channel-priority python=3.14 pyside6 ffmpeg "numcodecs<0.16" uv -y

On macOS, select the OpenBLAS build that avoids loading a second OpenMP runtime alongside PyTorch:

conda create -n das -c conda-forge --strict-channel-priority python=3.14 pyside6 ffmpeg "numcodecs<0.16" "libopenblas=*=*pthreads*" uv -y

Then activate the environment and install DAS:

conda activate das
uv pip install das --torch-backend=auto

--torch-backend=auto selects a suitable PyTorch build for the available hardware. To request a specific build instead, replace auto with one of these backends:

uv pip install das --torch-backend=cpu       # CPU only
uv pip install das --torch-backend=cu130     # NVIDIA CUDA 13.0
uv pip install das --torch-backend=rocm7.2   # AMD ROCm 7.2 on Linux
uv pip install das --torch-backend=xpu       # Intel GPU

Available accelerator builds depend on the operating system and hardware. On macOS, the standard PyTorch build supports Apple Metal acceleration.

Verify the installation

das version
das gui

Next steps

If all is working, you can now use DAS to annotate song. To get started, you will first need to train a network on your own data. For that you need annotated audio - either create new annotations using the GUI or convert existing annotations using python scripts.