Stable Audio 3 Medium โ GGUF (for sa3.cpp)
GGUF conversions of stabilityai/stable-audio-3-medium for sa3.cpp โ a portable C++/GGML port of Stable Audio 3, no PyTorch in the loop. Runs on CPU, CUDA, Vulkan, or Metal (Apple Silicon). Every component is validated against the PyTorch reference at cosine similarity ~1.0.
Files
This is a multi-file model. Grab the DiT + SAME at your chosen precision and the conditioner, plus the shared encoder + tokenizer from the t5gemma-b-b-ul2-GGUF repo.
| component | file | notes |
|---|---|---|
| DiT (diffusion transformer) | stable-audio-3-medium-dit-1.5B-v1.0-{F32,F16,Q8_0,Q5_K_M,Q4_K_M}.gguf |
pick one encoding |
| autoencoder (SAME-L) | stable-audio-3-medium-same-l-v1.0-{F32,F16,Q8_0,Q5_K_M,Q4_K_M}.gguf |
must match the DiT |
| conditioner | stable-audio-3-medium-conditioner-v1.0-F32.gguf |
tiny sidecar (prompt padding + seconds_total) |
| encoder + tokenizer | โ t5gemma-b-b-ul2-GGUF | shared across all SA3 variants |
F16 is the production path (~3.5s for 12s of audio on an 8GB laptop GPU); F32 is for CPU validation. The conditioner + encoder + tokenizer stay F32 (small / quality-critical).
Encodings
sa3-generate --encoding resolves the DiT and the SAME with the same suffix, so download the pair.
| encoding | DiT | SAME-L | total |
|---|---|---|---|
| F32 | 5545 MB | 3251 MB | 8796 MB |
| F16 | 2773 MB | 1626 MB | 4399 MB |
| Q8_0 | 1483 MB | 865 MB | 2348 MB |
| Q5_K_M | 1053 MB | 617 MB | 1670 MB |
| Q4_K_M | 962 MB | 570 MB | 1532 MB |
q4_k_m and q5_k_m promote the attention V, feed-forward down and embedding tensors to Q6_K;
q8_0 is uniform. Every tier passes sa3-quant-check with below-threshold=0 at cosine 0.990
against the F16 reference, for the DiT and the SAME alike.
Quantization buys footprint everywhere and speed only on some backends. CUDA and Vulkan gain roughly 33% end to end. Metal is flat โ Q8_0 is 1.7% faster and Q4_K_M 2.1% slower than F16, because the load-time saving and the added per-step dequant cancel out. On a Mac, pick a quant for the memory, not for the speed.
Usage
For use with sa3.cpp:
# pip install huggingface_hub
python tools/download_models.py --variant medium --encoding f16 # fetches this set + the shared encoder
# --model resolves the 5 gguf files in ./models by name
sa3-generate --model medium --prompt "upbeat funk groove with slap bass" --out song.wav
For a quantized set, pass the encoding to both โ the downloader and the generator use the same names:
python tools/download_models.py --variant medium --encoding q4_k_m
sa3-generate --model medium --encoding q4_k_m --prompt "upbeat funk groove with slap bass" --out song.wav
Performance
Roughly 3s for a 12s clip at f16 on an 8GB laptop GPU (RTX 5070), and ~6s on an Apple M4 โ end to end, including model load. The sliding-window decoder keeps long generations linear (a 2-minute clip is ~9s on the 5070). CPU works but is ~10ร slower. Full numbers + the f16 / flash-attention levers: docs/BENCHMARKS.md.
License
These are format conversions of stabilityai/stable-audio-3-medium, whose weights Stability AI releases under the Stability AI Community License: free for organizations under $1M annual revenue, with commercial use, fine-tuning, and derivative works permitted within that threshold (above it, contact Stability AI for an Enterprise License). Outputs are yours. That license carries over to these converted weights.
The upstream stable-audio-3 source code is released separately under MIT. Pair these with the shared T5Gemma text encoder, which is Google's under the Gemma Terms of Use.
Relationship to the original
Format conversions (weights โ GGUF) for inference in sa3.cpp โ no retraining, no architectural changes. See sa3.cpp/docs/DISTRIBUTION.md for the naming convention and how the pieces fit together.
- Downloads last month
- 710
4-bit
5-bit
8-bit
16-bit
32-bit
Model tree for thepatch/stable-audio-3-medium-GGUF
Base model
stabilityai/stable-audio-3-medium-base