stabilityai

/

stable-audio-open-1.0

like 950

Follow

Stability AI 5,891

+1

gradio

@NickyNicky
https://huggingface.co/spaces/ameerazam08/stableaudio-open-1.0

I'm not very experienced with Python but I could make it work like this:

=== requirements1.txt ===

-i https://download.pytorch.org/whl/cu121
torch
torchvision
torchaudio

=== requirements2.txt ===

einops
ninja
packaging
huggingface_hub
stable_audio_tools

=== infer.py ===

import torch
import torchaudio
from einops import rearrange
from stable_audio_tools import get_pretrained_model
from stable_audio_tools.inference.generation import generate_diffusion_cond

device = "cuda" if torch.cuda.is_available() else "cpu"

# Download model
model, model_config = get_pretrained_model("stabilityai/stable-audio-open-1.0")
sample_rate = model_config["sample_rate"]
sample_size = model_config["sample_size"]

model = model.to(device)

# Set up text and timing conditioning
conditioning = [{
    "prompt": "128 BPM dnb drums loop",
    "seconds_start": 0, 
    "seconds_total": 30
}]

# Generate stereo audio
output = generate_diffusion_cond(
    model,
    steps=100,
    cfg_scale=7,
    conditioning=conditioning,
    sample_size=sample_size,
    sigma_min=0.3,
    sigma_max=500,
    sampler_type="dpmpp-3m-sde",
    device=device,
    # force seed because of error on Win11 64 bits
    seed=123456
)

# Rearrange audio batch to a single sequence
output = rearrange(output, "b d n -> d (b n)")

# Peak normalize, clip, convert to int16, and save to file
output = output.to(torch.float32).div(torch.max(torch.abs(output))).clamp(-1, 1).mul(32767).to(torch.int16).cpu()
torchaudio.save("output.wav", output, sample_rate)

Then type

python -m venv venv
.\venv\Scripts\activate
pip install -r requirements1.txt
pip install -r requirements2.txt

(will take a while)
Then authenticate to your Hugging face account with a token (create token at https://huggingface.co/settings/tokens)

huggingface-cli login

(paste your token and follow the instructions, token will not be outputed when pasted)

Then

python .\infer.py

Edit infer.py as needed.
Change prompt and seed if necessary.

seed=123456

Edit infer.py as needed.
Change prompt and seed if necessary.

The seed doesnt work like that on windows, you will need to change line 138 in generation.py
"myenv/Lib/site-Packages/stable_audio_tools/inference/generation.py"
seed = 12345

change the seed in "myenv/Lib/site-Packages/stable_audio_tools/inference/generation.py"
from
seed = seed if seed != -1 else np.random.randint(0, 232 - 1)
to
seed = seed if seed != -1 else np.random.randint(0, 232 - 1,dtype=np.int64)
is also worked on my device.