Upload 2 files

README.md

@@ -4,22 +4,22 @@ license: apache-2.0
 
 # LocalSong
 
-LocalSong is an audio generation model focused on melodic instrumental music that uses tag-based conditioning to generate audio.
+LocalSong is a 700M parameter audio generation model focused on melodic instrumental music that uses tag-based conditioning.
 
 ## Installation
 
 ### Prerequisites
 
 - Python 3.10 or higher
-- CUDA-capable GPU recommended
+- CUDA-capable GPU recommended with 8GB of VRAM
 
 ### Setup
 
 git clone https://huggingface.co/Localsong/LocalSong
 cd localsong
-python3 -m venv venv
+python3.10 -m venv venv
 source venv/bin/activate
-pip install -r requirements.txt
+pip install torch==2.7.1 torchvision==0.22.1 torchaudio==2.7.1 --extra-index-url https://download.pytorch.org/whl/cu128
 
 ### Run
 
@@ -30,7 +30,10 @@ The interface will be available at `http://localhost:7860`
 ### Generation Advice
 
 Generations should use one of the soundtrack, soundtrack1 or soundtrack2 tags, as well as at least one other tag. They can use up to 8 tags; try combining genres and instruments.
+
 The default settings (CFG 3.5, steps 200) have been tested as optimal.
+If generation is too slow on your system, try lowering steps to 100.
+
 The first generation will be slower due to torch.compile, then speed will increase.
 The model was trained on vocals but not lyrics. Vocals will not have recognizable words.
 

requirements.txt

@@ -1,7 +1,6 @@
-torch>=2.8.0
-torchaudio>=2.8.0
-torchvision>=0.23.0
-torchcodec>=0.8.0
+torch==2.7.1
+torchaudio==2.7.1
+torchvision==0.22.1
 accelerate>=1.9.0
 diffusers>=0.34.0
 einops>=0.8.1