Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
34 commits
Select commit Hold shift + click to select a range
250baa8
Add XPU music generation pipeline and supporting files
rootLocalGhost Jan 27, 2026
7115859
Add XPU lyrics transcriber and improve downloader
rootLocalGhost Jan 30, 2026
910f3f5
Merge branch 'HeartMuLa:main' into main
rootLocalGhost Feb 7, 2026
466de53
docs: consolidate optimization docs into PERFORMANCE_OPTIMIZATION.md …
rootLocalGhost Feb 12, 2026
46c0edb
Delete echoes_of_silence.json
rootLocalGhost Feb 12, 2026
f2006af
Update .gitignore
rootLocalGhost Feb 12, 2026
da1cf62
Delete COMPLETE_OPTIMIZATION_SUMMARY.md
rootLocalGhost Feb 12, 2026
a0bef53
Delete PERFORMANCE_OPTIMIZATION.md
rootLocalGhost Feb 12, 2026
134de70
Create default.txt
rootLocalGhost Feb 13, 2026
4ceafa9
Delete echoes.txt
rootLocalGhost Feb 13, 2026
a436a91
Create lyrics.txt
rootLocalGhost Feb 13, 2026
6f38364
Create default.txt
rootLocalGhost Feb 13, 2026
94118fc
Delete sad_mood.txt
rootLocalGhost Feb 13, 2026
c190f5e
Create tags.txt
rootLocalGhost Feb 13, 2026
cb9739c
Update run.sh
rootLocalGhost Feb 13, 2026
16492b1
Update transcribe.sh
rootLocalGhost Feb 13, 2026
eb012d3
refactor(pipeline): add configurable parameters to HeartTranscriptorP…
rootLocalGhost Feb 14, 2026
f7d29ca
feat(cli): add compile_mode flag to transcribe.sh for torch.compile c…
rootLocalGhost Feb 14, 2026
cb0e7cc
perf(xpu): optimize music generation with torch.compile and memory ma…
rootLocalGhost Feb 14, 2026
c84bca3
feat(transcriber): enhance XPU transcription with torch.compile and t…
rootLocalGhost Feb 14, 2026
addff1b
docs(webui): add comprehensive guide for web interface setup
rootLocalGhost Feb 14, 2026
18305c1
docs(xpu): document Intel XPU setup and optimization procedures
rootLocalGhost Feb 14, 2026
32d2267
feat(examples): add LLM generator example for music creation
rootLocalGhost Feb 14, 2026
88a15da
feat(scripts): add web UI launcher for LLM-powered music generation
rootLocalGhost Feb 14, 2026
ebb1de0
feat(scripts): add web UI launcher script
rootLocalGhost Feb 14, 2026
1641dbe
build(deps): add Intel XPU specific dependencies
rootLocalGhost Feb 14, 2026
0c994af
feat(scripts): add optimized execution script for XPU hardware
rootLocalGhost Feb 14, 2026
a620d2e
feat(llm): add LLM module for music generation pipeline
rootLocalGhost Feb 14, 2026
feddd3e
feat(scripts): add optimized transcription script with XPU support
rootLocalGhost Feb 14, 2026
f931ab9
feat(validation): add XPU environment validation utility
rootLocalGhost Feb 14, 2026
805d62c
feat(webui): implement web interface for music generation
rootLocalGhost Feb 14, 2026
f344bb2
feat(webui): implement LLM-powered web interface for music generation
rootLocalGhost Feb 14, 2026
d0efb91
Merge branch 'HeartMuLa:main' into main
rootLocalGhost Mar 1, 2026
76c26aa
Merge branch 'main' into main
rootLocalGhost Mar 9, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -8,4 +8,5 @@ checkpoint/
**.egg-info/
build/
.idea/
ckpt/
ckpt/
output/
41 changes: 41 additions & 0 deletions HELP_GUIDES.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,41 @@
# HELP_GUIDES — Help, Guides & Troubleshooting (consolidated)

Purpose: single help document that combines quick reference, user guides and troubleshooting.

## Quick commands (one-liners)
- Validate setup: `python validate_xpu_setup.py`
- Generate (optimized): `./run_optimized.sh --lyrics ./inputs/lyrics/default.txt`
- Transcribe (optimized): `./transcribe_optimized.sh --music_path ./audio/song.mp3`
- Monitor GPU: `watch -n 0.5 'xpu-smi dump -m 0,1,5'` or `intel_gpu_top`

## Common flags explained
- `--cfg_scale` — 1.0 = speed, >1 = stronger guidance/quality
- `--max_audio_length_ms` — longer => better GPU utilization
- `--batch_size` (transcription) — larger = faster if VRAM permits
- `--chunk_length_s` (transcription) — longer chunks = better accuracy, more memory

## Typical user workflows
1. Quick test: `python xpu_music_gen.py --max_audio_length_ms 30000`
2. Full run with launcher: `./run_optimized.sh --lyrics my_lyrics.txt --save_path out.mp3`
3. Verify with transcription: `./transcribe_optimized.sh --music_path out.mp3 --save_path verified.txt`

## Troubleshooting (common problems & solutions)
- XPU not detected: verify PyTorch 2.10+ with XPU support and Level Zero drivers.
- Low GPU utilization: increase `--max_audio_length_ms`, check for CPU bottleneck (htop).
- Out of memory (OOM): reduce `--batch_size` or `--max_audio_length_ms`; set `--cfg_scale 1.0`.
- Slow checkpoint loading: ensure `transformers>=4.57.0` and SSD storage; use `low_cpu_mem_usage=True`.

## Transcription-specific tips
- Defaults: `--batch_size 16`, `--chunk_length_s 30`
- For speed: increase batch_size and lower chunk_length
- For accuracy: increase chunk_length and use cleaner audio

## When to open an issue
Include: output of `python validate_xpu_setup.py`, relevant command used, and `xpu-smi`/`intel_gpu_top` output.

## Where to find more detail
- Performance & tuning: `PERFORMANCE_OPTIMIZATION.md`
- Full original doc (backup): `COMPLETE_OPTIMIZATION_SUMMARY.md`

---
If you want I can add these links to `README.md` or remove the old split docs now.
Loading