add fast captioning module (CLAP + faster-whisper + Silero VAD), update deps

random 60s crop at training time (matches Side-Step chunk-duration), remove pre-split chunking

audio-level chunking (not latent), auto-scale epochs for chunk count

chunk latents into ~30s segments for faster CPU training, energy-aware boundaries

default rank 16 for faster CPU training

add credit lines to both tabs

compact Generate Music UI, same layout style as Train tab

move audio upload to left column

skip bare librosa sidecar, let preprocessing faf analysis handle caption fallback

fix adapter save path, smart LM fallback, compact training UI, remove Server Status

cancel, captioning, preprocessing, sidecar upload, elapsed time, GeneratorExit fix

fix review: debug leak, int crash, rank mismatch, 0-byte skip, log cap, understand diag

fix: save PEFT adapter (not full model), remove random suffix from LoRA names, fix epoch cap to 1000

remove XL checkpoint download (OOMKilled build, training uses standard turbo)

fix: save_every_n_epochs=0, add demucs-infer to Dockerfile, debug adapter dir

fix all review issues: dedup sampling/unwrap, thread-safe lock, cleanup, retry, security docs

update README with final state, full pipeline inference, LM generation step

fix inference: add LM generation step, detokenize codes before DiT, full pipeline working

add _is_space flag, block inference during training, understand clone fix

fix understand_audio: clone tensors for inference mode, working on GPU (52s)

add understand_audio (LM reverse), demucs-infer fix, commit refs, dtype fixes

major update: PyTorch inference, Gradio 6, session isolation, /understand captioning

truncate long files to fit cap, show which files truncated/skipped

accept files until total audio cap reached, skip rest with warning

replace per-file cap with total audio cap (30 min max)

add LoRA download button after training (gr.File output, like rvc-beatrice)

remove ace-server understand proxy, captioning stays librosa + txt sidecars

add /understand API endpoint for GGUF audio captioning

SDPA first on Blackwell, FA2 only for Ampere/Hopper, txt caption support

add GPU/CUDA auto-detect, mixed precision, flash_attn, txt caption parser

update defaults: LR 3e-4, rank 32, alpha 2x rank (per Side-Step author)

save only on cancel/finish, max epochs 1000, default 3

add mid/sas analysis modes (Demucs + ensemble), auto-select by dataset size

add auto-captioning (BPM/key/signature via librosa), add librosa+mutagen deps

switch training to standard turbo (11s/epoch), auto-select standard GGUF for LoRA inference

add XL variant mappings to _VARIANT_DIR

fix: train on XL turbo (matches XL GGUF for inference), add XL checkpoint download

add full README with API docs, MCP, CLI, architecture

allow custom LoRA values in dropdown (API clients)

log ace-server restart, show output for adapter debugging

fix: forward adapter to synth request, default LM to 1.7B

fix: adapter saved to clean dir, LM dropdown no 'Default', on-demand download

copy train_engine.py into Docker image

Side-Step training engine, tested locally on CPU

switch back to 1.7B LM (fastest at 269s, 0.6B was 936s)

swap LM to 0.6B Q8_0 for speed test

swap LM 4B->1.7B Q8_0 for faster CPU inference

add LoRA adapter dropdown to inference UI

default mp3, remove format selector, increase LM timeout to 900s

remove accelerate (causes meta tensors), clean up patches