Update to work as package

Author: julian-parker

README.md

@@ -73,7 +73,7 @@ The model is trained with utterances normalized to -20 LUFS. The `encode` functi
 
 ## Finetune
 
-To finetune a model given its config and checkpoint, execute `train.py` file. You can also directly use `stable-audio-tools`:
+To finetune a model given its config and checkpoint, execute `train.py` file:
 
 ```bash
 python train.py \

pyproject.toml

@@ -0,0 +1,3 @@
+[build-system]
+requires = ["setuptools","wheel"]
+build-backend = "setuptools.build_meta"

requirements.txt

@@ -0,0 +1,21 @@
+from setuptools import setup, find_packages
+
+setup(
+    name='stable-codec',
+    version='0.1.0',
+    author='Stability AI',
+    author_email='[email protected]',
+    description='Stable Codec: A series of codec models for speech and audio',
+    long_description=open('README.md').read(),
+    long_description_content_type='text/markdown',
+    url='https://github.com/Stability-AI/stable-codec/',
+    packages=find_packages(),
+    python_requires='>=3.10',
+    install_requires=['packaging',
+                      'wheel',
+                      'torch==2.4',
+                      'torchaudio==2.4',
+                      'stable-audio-tools==0.0.17',
+                      'pytorch-lightning==2.1',
+                      'prefigure==0.0.9']
+)

setup.py

@@ -0,0 +1 @@
+from stable_codec.model import StableCodec

stable_codec/__init__.py

@@ -11,7 +11,7 @@
 from stable_audio_tools.training.utils import copy_state_dict
 from stable_audio_tools.data.utils import VolumeNorm
 
-from residual_fsq import ResidualFSQBottleneck
+from .residual_fsq import ResidualFSQBottleneck
 from stable_audio_tools import get_pretrained_model
 
 class StableCodec(nn.Module):

ctc_loss.py renamed to stable_codec/ctc_loss.py

@@ -27,7 +27,7 @@
     create_optimizer_from_config, create_scheduler_from_config, log_metric,
 )
 
-from ctc_loss import CTCLossModule, PERModule
+from .ctc_loss import CTCLossModule, PERModule
 
 def trim_to_shortest(a, b):
     """Trim the longer of two tensors to the length of the shorter one."""

data/Text2Phone/Text2PhoneTokenizer.py renamed to stable_codec/data/Text2Phone/Text2PhoneTokenizer.py

@@ -9,9 +9,9 @@
 from stable_audio_tools.models.utils import load_ckpt_state_dict, remove_weight_norm_from_model
 from stable_audio_tools.training.utils import copy_state_dict
 
-from training_module import create_training_wrapper_from_config
-from training_demo import create_demo_callback_from_config
-from data.dataset import create_dataloader_from_config
+from stable_codec.training_module import create_training_wrapper_from_config
+from stable_codec.training_demo import create_demo_callback_from_config
+from stable_codec.data.dataset import create_dataloader_from_config
 
 class ExceptionCallback(pl.Callback):
     def on_exception(self, trainer, module, err):