AnchorSteer Weights

--- license: mit tags: - audio - music - music-editing - diffusion - stable-audio pipeline_tag: audio-to-audio --- # AnchorSteer Pretrained Weights Pretrained Controller checkpoints for the KDD 2026 paper: **AnchorSteer: Self-Discovered Concept Injection for Structure-Preserving Music Editing** - GitHub: [hengtsune1024/AnchorSteer](https://github.com/hengtsune1024/AnchorSteer) - Code archive: [10.5281/zenodo.20368360](https://doi.org/10.5281/zenodo.20368360) --- ## Contents Each checkpoint directory contains: | File | Description | |------|-------------| | `best.pth` | Controller weights at the epoch with lowest validation loss | | `config.yaml` | Full training configuration for reproducibility | This release includes 5 concepts: | Directory | Concept type | Concept | |-----------|-------------|---------| | `exps/ambient-transformer/` | Genre | Ambient | | `exps/harp-transformer/` | Instrument | Harp | | `exps/jazz-transformer/` | Genre | Jazz | | `exps/piano-transformer/` | Instrument | Piano | | `exps/rock-transformer/` | Genre | Rock | All checkpoints use the **Transformer** Controller architecture trained on full 47-second audio (20 epochs, lr=5e-4, DiT blocks 0–23). --- ## Usage ### 1. Install AnchorSteer Follow the [installation instructions](https://github.com/hengtsune1024/AnchorSteer#installation) in the GitHub repo. ### 2. Download checkpoints ```bash # from the AnchorSteer root directory hf download heng1024/AnchorSteer-weights --repo-type model --include "exps/*" --local-dir . ``` This places the `exps/` folder directly into the repo root, matching the expected layout. ### 3. Run inference ```bash # Generate concept-steered audio from a text prompt python generate.py --exp_dir exps/rock-transformer # Structure-preserving editing of a source file python edit.py --source_audio path/to/source.wav --concept_dir exps/rock-transformer ``` See the [README](https://github.com/hengtsune1024/AnchorSteer#quick-start-inference-with-pre-trained-checkpoints) for full options. --- ## Citation ```bibtex @inproceedings{anchosteer2026, title = {AnchorSteer: Self-Discovered Concept Injection for Structure-Preserving Music Editing}, author = {Chih-Heng Chang, Keng-Seng Ho, Chih-Yu Tsai, Kuan-Lin Chen, Yi-Hsuan Yang, Jian-Jiun Ding}, booktitle = {Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V...

Source context: 0 downloads · 0 likes · Pipeline audio-to-audio · Repo heng1024/AnchorSteer-weights