HOT-Step-CPP StableStep Models
==============================

This repository contains format-converted derivatives (ONNX and GGUF) of:

  stabilityai/stable-audio-3-medium
  https://huggingface.co/stabilityai/stable-audio-3-medium

comprising the Stable Audio 3 medium diffusion transformer, the SAME-L
audio autoencoder, and the T5Gemma text encoder (originally by Google,
distributed as part of the Stable Audio 3 release), plus the associated
tokenizer files.

No fine-tuning was performed. Modifications are limited to:
  - export to ONNX graph format (fp32)
  - conversion to GGUF (BF16/F16 weights, F32 for precision-sensitive tensors)
  - splitting into encoder/decoder/text-encoder components
  - dropping the unused T5Gemma decoder half

These derivatives are distributed under the Stability AI Community License:
  https://stability.ai/community-license-agreement

Required attribution (retained per Section IV(a) of the license):

  "This Stability AI Model is licensed under the Stability AI Community
   License, Copyright (c) Stability AI Ltd. All Rights Reserved"

Powered by Stability AI.