ACE-Step/inference.ipynb

ACE-Step Inference

A Step Towards Music Generation Foundation Model

Credits:

• Ace-Step by Ace-Step

• Colab improvement by NeoDev

🖥️ Hardware Performance

We have evaluated ACE-Step across different hardware setups, yielding the following throughput results:

Device	RTF (27 steps)	Time to render 1 min audio (27 steps)	RTF (60 steps)	Time to render 1 min audio (60 steps)
NVIDIA RTX 4090	34.48 ×	1.74 s	15.63 ×	3.84 s
NVIDIA A100	27.27 ×	2.20 s	12.27 ×	4.89 s
NVIDIA RTX 3090	12.76 ×	4.70 s	6.48 ×	9.26 s
MacBook M2 Max	2.27 ×	26.43 s	1.03 ×	58.25 s

We use RTF (Real-Time Factor) to measure the performance of ACE-Step. Higher values indicate faster generation speed. 27.27x means to generate 1 minute of music, it takes 2.2 seconds (60/27.27). The performance is measured on a single GPU with batch size 1 and 27 steps.

In [ ]:

#@title Install and Download


import codecs



print("Installing...")
!sudo apt update > /dev/null 2>&1
!sudo apt install python3.10 > /dev/null 2>&1
!sudo update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.10 1 > /dev/null 2>&1
!sudo update-alternatives --set python3 /usr/bin/python3.10 > /dev/null 2>&1
!curl -sS https://bootstrap.pypa.io/get-pip.py | python3 > /dev/null 2>&1
import sys
sys.path.append('/usr/local/lib/python3.10/dist-packages')


repopath = codecs.decode('erdhverzragf.gkg', 'rot_13')


!git clone https://github.com/ace-step/ACE-Step
%cd /content/ACE-Step


!pip install uv pyngrok > /dev/null 2>&1
!uv pip install -r {repopath} > /dev/null 2>&1
!uv pip install huggingface-hub numpy==1.26.0 > /dev/null 2>&1
!huggingface-cli download ACE-Step/ACE-Step-v1-3.5B --local-dir /content/ACE-Step/checkpoints --local-dir-use-symlinks False


import os
os.environ['MPLBACKEND'] = 'agg'

In [ ]:

#@title Run Gradio UI
bf16 = True # @param {"type":"boolean"}


print(" * Running UI...")

!pip install e .!acestep --checkpoint_path ./checkpoints/ --port 7865 --device_id 0 --share true --bf16 {bf16}