Sync to HuggingFace Spaces

.gitattributes

@@ -0,0 +1,2 @@
+# Auto detect text files and perform LF normalization
+* text=auto

.github/workflows/sync.yml

@@ -0,0 +1,26 @@
+name: Sync to Hugging Face Spaces
+
+on:
+    push:
+        branches:
+            - main
+
+jobs:
+    sync:
+        name: Sync
+        runs-on: ubuntu-latest
+
+        steps:
+            - name: Checkout Repository
+              uses: actions/checkout@v4
+              with:
+                  lfs: true
+
+            - name: Sync to Hugging Face Spaces
+              uses: JacobLinCool/huggingface-sync@v1
+              with:
+                  github: ${{ secrets.GITHUB_TOKEN }}
+                  user: jacoblincool # Hugging Face username or organization name
+                  space: audio-super-resolution # Hugging Face space name
+                  token: ${{ secrets.HF_TOKEN }} # Hugging Face token
+                  configuration: headers.yaml

.gitignore

@@ -0,0 +1,160 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/#use-with-ide
+.pdm.toml
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024 JacobLinCool
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -0,0 +1,15 @@
+---
+title: Audio Super Resolution
+emoji: 🎧
+colorFrom: red
+colorTo: gray
+sdk: gradio
+sdk_version: 4.37.2
+app_file: app.py
+pinned: false
+license: mit
+---
+
+# Audio Super Resolution
+
+Audio Super Resolution.

app.py

@@ -0,0 +1,138 @@
+import os
+from typing import List
+import gradio as gr
+import tempfile
+import numpy as np
+import librosa
+import soundfile as sf
+import spaces
+from audiosr import super_resolution
+from audiosr_utils import load_audiosr
+
+
+audiosr_model = load_audiosr()
+
+
+def split_audio_to_chunks(y, sr=48000, chunk_duration=5.12) -> List[str]:
+    # Calculate the number of samples per chunk
+    chunk_samples = int(chunk_duration * sr)
+
+    # Split the audio into chunks
+    chunks = [y[i : i + chunk_samples] for i in range(0, len(y), chunk_samples)]
+
+    # Save each chunk to a temporary file
+    temp_files = []
+    for i, chunk in enumerate(chunks):
+        # Create a temporary file
+        temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
+        temp_files.append(temp_file.name)
+
+        # Write the chunk to the temporary file
+        sf.write(temp_file.name, chunk, sr)
+
+    return temp_files
+
+
+@spaces.GPU(duration=180)
+def run_audiosr(
+    chunks: List[str], guidance_scale: float, ddim_steps: int
+) -> np.ndarray:
+    waveforms = []
+    for i, chunk in enumerate(chunks):
+        print(f"Processing chunk {i+1}/{len(chunks)}")
+        waveform = super_resolution(
+            audiosr_model,
+            chunk,
+            guidance_scale=guidance_scale,
+            ddim_steps=ddim_steps,
+        )
+        waveforms.append(waveform)
+    waveform = np.concatenate(waveforms, axis=-1)  # (1, 1, N)
+    waveform = waveform.squeeze()
+    return waveform
+
+
+def audiosr_infer(audio: str) -> str:
+    guidance_scale = 3.5
+    ddim_steps = 100
+
+    y, sr = librosa.load(audio, sr=48000)
+    if len(y) > 60 * sr:
+        y = y[: 60 * sr]
+        gr.Info("Audio is too long, only the first 60 seconds will be processed")
+
+    chunk_files = split_audio_to_chunks(y, sr=sr, chunk_duration=5.12)
+    print(f"Splited audio chunks: {chunk_files}")
+
+    waveform = run_audiosr(chunk_files, guidance_scale, ddim_steps)
+    sr = 44100
+
+    for chunk_file in chunk_files:
+        os.remove(chunk_file)
+
+    with tempfile.NamedTemporaryFile(suffix=".mp3", delete=False) as f:
+        sf.write(f.name, waveform, sr)
+        return f.name
+
+
+models = {
+    "AudioSR": audiosr_infer,
+}
+
+
+def infer(audio: str, model: str, sr: int) -> str:
+    if sr > 0:
+        # resample audio
+        y, _ = librosa.load(audio, sr=sr)
+        # save resampled audio
+        with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as f:
+            sf.write(f.name, y, sr)
+            return models[model](f.name)
+    else:
+        return models[model](audio)
+
+
+with gr.Blocks() as app:
+    with open(os.path.join(os.path.dirname(__file__), "README.md"), "r") as f:
+        README = f.read()
+        # remove yaml front matter
+        blocks = README.split("---")
+        if len(blocks) > 1:
+            README = "---".join(blocks[2:])
+
+    gr.Markdown(README)
+
+    with gr.Row():
+        with gr.Column():
+            gr.Markdown("## Upload an audio file")
+            audio = gr.Audio(label="Upload an audio file", type="filepath")
+            sr = gr.Slider(
+                value=0,
+                label="Resample audio to sample rate before inference, 0 means no resampling",
+                minimum=0,
+                maximum=48000,
+                step=1000,
+            )
+
+    with gr.Row():
+        model = gr.Radio(
+            label="Select a model",
+            choices=[s for s in models.keys()],
+            value="AudioSR",
+        )
+        btn = gr.Button("Infer")
+
+    with gr.Row():
+        with gr.Column():
+            out = gr.Audio(
+                label="Output", format="mp3", type="filepath", interactive=False
+            )
+
+    btn.click(
+        fn=infer,
+        inputs=[audio, model, sr],
+        outputs=[out],
+        api_name="infer",
+    )
+
+app.launch(show_error=True)

audiosr_utils.py

@@ -0,0 +1,42 @@
+import torch
+import yaml
+from audiosr import download_checkpoint, default_audioldm_config, LatentDiffusion
+
+
+def load_audiosr(ckpt_path=None, config=None, device=None, model_name="basic"):
+    if device is None or device == "auto":
+        if torch.cuda.is_available():
+            device = torch.device("cuda:0")
+        elif torch.backends.mps.is_available():
+            device = torch.device("mps")
+        else:
+            device = torch.device("cpu")
+
+    print("Loading AudioSR: %s" % model_name)
+    print("Loading model on %s" % device)
+
+    ckpt_path = download_checkpoint(model_name)
+
+    if config is not None:
+        assert type(config) is str
+        config = yaml.load(open(config, "r"), Loader=yaml.FullLoader)
+    else:
+        config = default_audioldm_config(model_name)
+
+    # # Use text as condition instead of using waveform during training
+    config["model"]["params"]["device"] = device
+    # config["model"]["params"]["cond_stage_key"] = "text"
+
+    # No normalization here
+    latent_diffusion = LatentDiffusion(**config["model"]["params"])
+
+    resume_from_checkpoint = ckpt_path
+
+    checkpoint = torch.load(resume_from_checkpoint, map_location="cpu")
+
+    latent_diffusion.load_state_dict(checkpoint["state_dict"], strict=True)
+
+    latent_diffusion.eval()
+    latent_diffusion = latent_diffusion.to(device)
+
+    return latent_diffusion

headers.yaml

@@ -0,0 +1,9 @@
+title: Audio Super Resolution
+emoji: 🎧
+colorFrom: red
+colorTo: gray
+sdk: gradio
+sdk_version: 4.37.2
+app_file: app.py
+pinned: false
+license: mit

requirements.txt

@@ -0,0 +1,7 @@
+gradio
+numpy
+librosa
+soundfile
+spaces
+huggingface_hub
+audiosr