Configuration Parsing Warning:In adapter_config.json: "peft.task_type" must be a string

Grunge

Grunge-style generation. Covers: 90s grunge rock, distorted power chords, raspy vocals, driving drums; heavy grunge, fuzzy bass, screamed vocals, dynamic loud-quiet structure; and melodic grunge, jangly clean guitars to distorted choruses, emotional vocals.

Quick start

  • Base model: AceStep v1.5 Turbo (2B)
  • Trigger word: roti-grungg (must appear in the prompt to activate)
  • Recommended strength: 1.0
  • Tested strength range: 0.3 โ€“ 1.0
  • Recommended steps: 8
  • Recommended guidance: 7.0
  • Recommended shift: 3.0
  • Sample rate: 48000 Hz

Example prompts

  • 90s grunge rock, distorted power chords, raspy vocals, driving drums (130 BPM, F# minor)
  • heavy grunge, fuzzy bass, screamed vocals, dynamic loud-quiet structure (120 BPM, D minor)
  • melodic grunge, jangly clean guitars to distorted choruses, emotional vocals (115 BPM, G minor)

Training

  • Rank 64, alpha 128, dropout 0.1
  • Target modules: q_proj, k_proj, v_proj, o_proj
  • 500 epochs / 7500 steps, batch 1, optimizer adamw_fused, lr 0.0003, warmup 100
  • Dataset: 15 synthetic tracks, 3712s total, instrumental fraction 0.0

Files

  • grunge-v1.safetensors โ€” PEFT LoRA weights (88,130,248 bytes, sha256 3e382f2fbb9bb6a2cdaabee2472e85baa627795a552be14befda4e0834283106)
  • grunge-v1.metadata.json โ€” machine-readable sidecar (schema v1)
  • adapter_config.json โ€” PEFT adapter config

Full metadata: see grunge-v1.metadata.json.

Downloads last month
8
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support