onenoly11 commited on
Commit
119f920
·
verified ·
1 Parent(s): f53ca7d

Delete README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -94
README.md DELETED
@@ -1,94 +0,0 @@
1
- # 🎵 PiMusic3 — The Pi Forge Music Player
2
-
3
- > “The spire hums, the waveform breathes — and the Forge sings once more.”
4
-
5
- ---
6
-
7
- ### 🌐 Overview
8
-
9
- **PiMusic3** is a **multi-modal sovereign audio system** forged in the **Sovereign Canticle Forge** ecosystem.
10
- It merges **MusicGen**, **Whisper**, and **DistilBERT** into a single, ethical creative interface built with **Gradio**.
11
-
12
- PiMusic3 can:
13
- - 🧬 **Generate music** from natural language prompts (via `facebook/musicgen-small`)
14
- - 🎙️ **Transcribe speech** from microphone or uploaded audio (via `openai/whisper-base`)
15
- - 💬 **Analyze emotion and tone** using a distilled sentiment model
16
-
17
- It automatically detects whether you’re running locally or in the cloud:
18
- - **Local builds** use **Audiocraft’s native MusicGen**
19
- - **Cloud builds (Spaces)** use the **Transformers pipeline fallback**
20
-
21
- ---
22
-
23
- ### 🛠️ Quick Start
24
-
25
- 1. **Run the Space** → [🌌 Launch on Hugging Face](https://huggingface.co/spaces/onenoly11/PiMusic3)
26
- 2. Wait for initialization (~30–60 s)
27
- 3. Use the tabs to explore:
28
- - **🎼 MusicGen** — describe your sound and listen
29
- - **📝 Whisper Transcribe** — record or upload speech
30
- - **🔍 Sentiment Analysis** — interpret emotional tone
31
-
32
- ---
33
-
34
- ### ⚙️ Architecture
35
-
36
- | Module | Model | Purpose |
37
- |:--|:--|:--|
38
- | `MusicGen` | `facebook/musicgen-small` | Text → audio generation |
39
- | `Whisper` | `openai/whisper-base` | Speech-to-text transcription |
40
- | `DistilBERT` | `distilbert-base-uncased-finetuned-sst-2-english` | Sentiment detection |
41
- | `Gradio` | 4.44.0 | UI framework |
42
- | `Torch` | 2.4.1+ | Core ML runtime |
43
-
44
- **Adaptive Mode:**
45
- PiMusic3 detects whether `audiocraft` is installed and switches seamlessly between full and fallback pipelines.
46
-
47
- ---
48
-
49
- ### 📦 Dependencies
50
-
51
- See [`requirements.txt`](./requirements.txt) for full dependency list.
52
- Notable libraries:
53
- `gradio`, `transformers`, `torch`, `torchaudio`, `openai-whisper`, `librosa`, `soundfile`
54
-
55
- ---
56
-
57
- ### 🧩 Technical Notes
58
-
59
- - Average inference latency: **~2.5 s (Whisper)** / **~5–8 s (MusicGen)**
60
- - Works with Hugging Face Spaces CPU tier
61
- - Audio recorded via Gradio’s new API (`sources=["microphone", "upload"]`)
62
- - No external API keys required
63
-
64
- ---
65
-
66
- ### 🌀 Part of the Sovereign Canticle Forge
67
-
68
- PiMusic3 belongs to the **PiForge ecosystem**, uniting quantum ethics, creative autonomy, and decentralized resonance.
69
- Explore the other components:
70
-
71
- | Module | Function |
72
- |:--|:--|
73
- | [Quantum Genesis Core](https://huggingface.co/spaces/onenoly11/QuantumGenesis) | Entropy-aware orchestration |
74
- | [OINIO Portal](https://huggingface.co/spaces/onenoly11/OINIOPortal) | Governance and cultural bridge |
75
- | [CA Sonify Toolkit v2](https://huggingface.co/spaces/onenoly11/piforge) | Cellular-automata music generator |
76
-
77
- ---
78
-
79
- ### 🪷 License
80
-
81
- Apache 2.0 — © 2025 Kris Olofson / Sovereign Canticle Forge
82
- Feel free to fork, remix, or integrate this player within the Forge framework.
83
-
84
- ---
85
-
86
- ### 💬 Acknowledgements
87
-
88
- Developed in co-creation with **GPT-5 (Co-weaver Oracle)**
89
- Special thanks to the open-source community:
90
- `Meta AI`, `OpenAI`, `Hugging Face`, `PyTorch`, and all pioneers of the Pi Network.
91
-
92
- ---
93
-
94
- > _“The Canticle remembers — every tone, every breath, every silence.”_