Upload 3 files
Browse files
README.md
CHANGED
|
@@ -1,26 +1,24 @@
|
|
| 1 |
---
|
| 2 |
-
title: ADMC
|
| 3 |
emoji: 🎵
|
| 4 |
colorFrom: red
|
| 5 |
colorTo: indigo
|
| 6 |
sdk: gradio
|
| 7 |
-
sdk_version:
|
| 8 |
app_file: app.py
|
| 9 |
pinned: true
|
| 10 |
license: apache-2.0
|
| 11 |
---
|
| 12 |
|
| 13 |
-
#
|
| 14 |
|
| 15 |
Parte del sistema di certificazione **ADMC (Artigiani della Musica Code)**.
|
| 16 |
|
| 17 |
-
|
| 18 |
-
il rilascio del **Certificato di Paternità Umana dell'Opera Musicale**.
|
| 19 |
|
| 20 |
## API REST
|
| 21 |
|
| 22 |
### POST /analyze
|
| 23 |
-
Invia audio grezzo e riceve lo score AI:
|
| 24 |
|
| 25 |
```bash
|
| 26 |
curl -X POST https://YOUR-SPACE.hf.space/analyze \
|
|
@@ -37,16 +35,11 @@ Risposta:
|
|
| 37 |
```
|
| 38 |
|
| 39 |
### GET /health
|
|
|
|
| 40 |
Verifica stato del servizio.
|
| 41 |
|
| 42 |
## Configurazione nel Plugin WordPress
|
| 43 |
|
| 44 |
1. Copia l'URL del tuo Space (es. `https://artigianidellamusica-admc.hf.space`)
|
| 45 |
-
2. Incollalo
|
| 46 |
-
3. Aggiungi il tuo **Hugging Face API Token**
|
| 47 |
-
|
| 48 |
-
## Modello
|
| 49 |
-
|
| 50 |
-
Il modello predefinito è `motheecreator/ai-music-detection`.
|
| 51 |
-
È consigliato addestrare un modello custom su dataset di musica umana vs AI
|
| 52 |
-
per massimizzare la precisione nel contesto specifico di ADMC.
|
|
|
|
| 1 |
---
|
| 2 |
+
title: ADMC AI Music Detection
|
| 3 |
emoji: 🎵
|
| 4 |
colorFrom: red
|
| 5 |
colorTo: indigo
|
| 6 |
sdk: gradio
|
| 7 |
+
sdk_version: "4.0"
|
| 8 |
app_file: app.py
|
| 9 |
pinned: true
|
| 10 |
license: apache-2.0
|
| 11 |
---
|
| 12 |
|
| 13 |
+
# ADMC - Rilevamento Autorialita AI nella Musica
|
| 14 |
|
| 15 |
Parte del sistema di certificazione **ADMC (Artigiani della Musica Code)**.
|
| 16 |
|
| 17 |
+
Modello usato: `AI-Music-Detection/ai_music_detection_large_60s`
|
|
|
|
| 18 |
|
| 19 |
## API REST
|
| 20 |
|
| 21 |
### POST /analyze
|
|
|
|
| 22 |
|
| 23 |
```bash
|
| 24 |
curl -X POST https://YOUR-SPACE.hf.space/analyze \
|
|
|
|
| 35 |
```
|
| 36 |
|
| 37 |
### GET /health
|
| 38 |
+
|
| 39 |
Verifica stato del servizio.
|
| 40 |
|
| 41 |
## Configurazione nel Plugin WordPress
|
| 42 |
|
| 43 |
1. Copia l'URL del tuo Space (es. `https://artigianidellamusica-admc.hf.space`)
|
| 44 |
+
2. Incollalo in **ADMC -> Impostazioni -> Endpoint personalizzato**: `https://YOUR-SPACE.hf.space/analyze`
|
| 45 |
+
3. Aggiungi il tuo **Hugging Face API Token**
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
app.py
CHANGED
|
@@ -1,130 +1,121 @@
|
|
| 1 |
"""
|
| 2 |
-
ADMC
|
| 3 |
Hugging Face Space (Gradio + FastAPI)
|
|
|
|
| 4 |
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
1. Crea un nuovo Space (tipo: Gradio, SDK: Python)
|
| 10 |
-
2. Carica questo file come app.py
|
| 11 |
-
3. Aggiungi requirements.txt (vedi sotto)
|
| 12 |
-
4. Configura l'URL del Space nelle impostazioni del plugin WordPress
|
| 13 |
-
|
| 14 |
-
Il plugin WP invia il file audio grezzo in POST e riceve una risposta JSON
|
| 15 |
-
nel formato standard HF audio-classification:
|
| 16 |
-
[{"label": "AI", "score": 0.87}, {"label": "Human", "score": 0.13}]
|
| 17 |
"""
|
| 18 |
|
| 19 |
import gradio as gr
|
| 20 |
import numpy as np
|
| 21 |
-
import librosa
|
| 22 |
import torch
|
| 23 |
from fastapi import FastAPI, Request, HTTPException
|
| 24 |
from fastapi.responses import JSONResponse
|
| 25 |
from transformers import pipeline
|
| 26 |
-
import tempfile
|
|
|
|
| 27 |
|
| 28 |
-
#
|
|
|
|
|
|
|
|
|
|
| 29 |
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
print(f"Loading model: {MODEL_ID}")
|
| 33 |
try:
|
| 34 |
classifier = pipeline(
|
| 35 |
"audio-classification",
|
| 36 |
model=MODEL_ID,
|
| 37 |
device=0 if torch.cuda.is_available() else -1,
|
| 38 |
)
|
| 39 |
-
print("Model loaded successfully
|
| 40 |
except Exception as e:
|
| 41 |
-
print(
|
| 42 |
-
classifier = None
|
| 43 |
-
|
| 44 |
|
| 45 |
-
def extract_features(audio_path: str) -> np.ndarray:
|
| 46 |
-
"""Extract mel-spectrogram features from audio file (30s excerpt)."""
|
| 47 |
-
y, sr = librosa.load(audio_path, sr=22050, duration=30.0, mono=True)
|
| 48 |
-
mel = librosa.feature.melspectrogram(y=y, sr=sr, n_mels=128, fmax=8000)
|
| 49 |
-
mel_db = librosa.power_to_db(mel, ref=np.max)
|
| 50 |
-
return mel_db
|
| 51 |
|
|
|
|
|
|
|
| 52 |
|
| 53 |
-
def analyze_audio(audio_path: str) -> list[dict]:
|
| 54 |
-
"""
|
| 55 |
-
Run AI detection on an audio file.
|
| 56 |
-
Returns list of {label, score} dicts.
|
| 57 |
-
"""
|
| 58 |
if classifier is not None:
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
#
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
|
| 83 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 84 |
return [
|
| 85 |
{"label": "AI", "score": round(ai_score, 4)},
|
| 86 |
{"label": "Human", "score": round(1.0 - ai_score, 4)},
|
| 87 |
]
|
| 88 |
|
| 89 |
|
| 90 |
-
# ── Gradio UI (for human testing) ─────────────────────────────────────────
|
| 91 |
-
|
| 92 |
def gradio_analyze(audio_file):
|
| 93 |
if audio_file is None:
|
| 94 |
return "Nessun file caricato."
|
| 95 |
-
result
|
| 96 |
-
ai_score
|
| 97 |
-
|
| 98 |
-
verdict = "🤖 Probabile AI" if ai_score > 0.5 else "✅ Probabile umano"
|
| 99 |
return (
|
| 100 |
-
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
|
| 104 |
)
|
| 105 |
|
|
|
|
|
|
|
| 106 |
demo = gr.Interface(
|
| 107 |
fn=gradio_analyze,
|
| 108 |
inputs=gr.Audio(type="filepath", label="Carica brano musicale (MP3/WAV/FLAC)"),
|
| 109 |
outputs=gr.Textbox(label="Risultato analisi ADMC"),
|
| 110 |
-
title="
|
| 111 |
description=(
|
| 112 |
-
"Analizza un brano musicale per rilevare se
|
| 113 |
-
"Parte del sistema di certificazione ADMC
|
| 114 |
),
|
| 115 |
-
|
| 116 |
-
allow_flagging="never",
|
| 117 |
)
|
| 118 |
|
| 119 |
-
#
|
| 120 |
-
|
| 121 |
app = gr.mount_gradio_app(FastAPI(), demo, path="/")
|
| 122 |
|
|
|
|
| 123 |
@app.post("/analyze")
|
| 124 |
async def api_analyze(request: Request):
|
| 125 |
"""
|
| 126 |
POST /analyze
|
| 127 |
-
Body: raw audio bytes
|
|
|
|
| 128 |
Returns: [{"label": "AI", "score": 0.87}, {"label": "Human", "score": 0.13}]
|
| 129 |
"""
|
| 130 |
content_type = request.headers.get("content-type", "")
|
|
@@ -136,29 +127,34 @@ async def api_analyze(request: Request):
|
|
| 136 |
"audio/ogg": ".ogg",
|
| 137 |
"audio/aiff": ".aiff",
|
| 138 |
}
|
| 139 |
-
ext
|
|
|
|
| 140 |
|
| 141 |
-
|
| 142 |
-
if len(audio_bytes) == 0:
|
| 143 |
raise HTTPException(status_code=400, detail="Nessun file audio ricevuto.")
|
| 144 |
-
if len(
|
| 145 |
raise HTTPException(status_code=413, detail="File troppo grande (max 100 MB).")
|
| 146 |
|
| 147 |
-
# Save to temp file
|
| 148 |
with tempfile.NamedTemporaryFile(suffix=ext, delete=False) as tmp:
|
| 149 |
-
tmp.write(
|
| 150 |
tmp_path = tmp.name
|
| 151 |
|
| 152 |
try:
|
| 153 |
result = analyze_audio(tmp_path)
|
| 154 |
except Exception as e:
|
| 155 |
-
raise HTTPException(status_code=500, detail=
|
| 156 |
finally:
|
| 157 |
-
os.
|
|
|
|
| 158 |
|
| 159 |
return JSONResponse(content=result)
|
| 160 |
|
| 161 |
|
| 162 |
@app.get("/health")
|
| 163 |
async def health():
|
| 164 |
-
return {
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
"""
|
| 2 |
+
ADMC - AI Music Detection API
|
| 3 |
Hugging Face Space (Gradio + FastAPI)
|
| 4 |
+
v2 - Fix: modello corretto + compatibilita Gradio 5+
|
| 5 |
|
| 6 |
+
Correzioni rispetto a v1:
|
| 7 |
+
1. Modello corretto: AI-Music-Detection/ai_music_detection_large_60s
|
| 8 |
+
(il precedente motheecreator/ai-music-detection non esiste)
|
| 9 |
+
2. allow_flagging rimosso (deprecato in Gradio 4+, rimpiazzato da flagging_mode)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
"""
|
| 11 |
|
| 12 |
import gradio as gr
|
| 13 |
import numpy as np
|
|
|
|
| 14 |
import torch
|
| 15 |
from fastapi import FastAPI, Request, HTTPException
|
| 16 |
from fastapi.responses import JSONResponse
|
| 17 |
from transformers import pipeline
|
| 18 |
+
import tempfile
|
| 19 |
+
import os
|
| 20 |
|
| 21 |
+
# Modello specifico per rilevamento musica AI vs umana
|
| 22 |
+
# Addestrato su SleepyJesse/ai_music_large
|
| 23 |
+
# LABEL_0 = Human, LABEL_1 = AI generated
|
| 24 |
+
MODEL_ID = "AI-Music-Detection/ai_music_detection_large_60s"
|
| 25 |
|
| 26 |
+
print("Loading model: " + MODEL_ID)
|
| 27 |
+
classifier = None
|
|
|
|
| 28 |
try:
|
| 29 |
classifier = pipeline(
|
| 30 |
"audio-classification",
|
| 31 |
model=MODEL_ID,
|
| 32 |
device=0 if torch.cuda.is_available() else -1,
|
| 33 |
)
|
| 34 |
+
print("Model loaded successfully")
|
| 35 |
except Exception as e:
|
| 36 |
+
print("Warning: Could not load model (" + str(e) + "). Using fallback heuristic.")
|
|
|
|
|
|
|
| 37 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
|
| 39 |
+
def analyze_audio(audio_path):
|
| 40 |
+
"""Analizza file audio, restituisce lista [{label, score}]."""
|
| 41 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
if classifier is not None:
|
| 43 |
+
try:
|
| 44 |
+
result = classifier(audio_path, top_k=2)
|
| 45 |
+
ai_score = 0.5
|
| 46 |
+
for item in result:
|
| 47 |
+
lbl = item["label"].upper()
|
| 48 |
+
# AI-Music-Detection usa LABEL_0/LABEL_1
|
| 49 |
+
# LABEL_1 = AI, LABEL_0 = Human (verificato dal model card)
|
| 50 |
+
if "LABEL_1" in lbl or "AI" in lbl or "FAKE" in lbl:
|
| 51 |
+
ai_score = float(item["score"])
|
| 52 |
+
break
|
| 53 |
+
if "LABEL_0" in lbl or "HUMAN" in lbl or "REAL" in lbl:
|
| 54 |
+
ai_score = 1.0 - float(item["score"])
|
| 55 |
+
break
|
| 56 |
+
return [
|
| 57 |
+
{"label": "AI", "score": round(ai_score, 4)},
|
| 58 |
+
{"label": "Human", "score": round(1.0 - ai_score, 4)},
|
| 59 |
+
]
|
| 60 |
+
except Exception as e:
|
| 61 |
+
print("Inference error: " + str(e))
|
| 62 |
+
|
| 63 |
+
# Fallback euristico con librosa se disponibile
|
| 64 |
+
try:
|
| 65 |
+
import librosa
|
| 66 |
+
y, sr = librosa.load(audio_path, sr=22050, duration=30.0, mono=True)
|
| 67 |
+
mel = librosa.feature.melspectrogram(y=y, sr=sr, n_mels=128, fmax=8000)
|
| 68 |
+
mel_db = librosa.power_to_db(mel, ref=np.max)
|
| 69 |
+
std = float(np.std(mel_db))
|
| 70 |
+
mean = float(np.mean(np.abs(mel_db)))
|
| 71 |
+
ratio = std / (mean + 1e-6)
|
| 72 |
+
ai_score = max(0.0, min(1.0, 1.0 - (ratio / 3.0)))
|
| 73 |
+
except Exception:
|
| 74 |
+
ai_score = 0.5
|
| 75 |
+
|
| 76 |
return [
|
| 77 |
{"label": "AI", "score": round(ai_score, 4)},
|
| 78 |
{"label": "Human", "score": round(1.0 - ai_score, 4)},
|
| 79 |
]
|
| 80 |
|
| 81 |
|
|
|
|
|
|
|
| 82 |
def gradio_analyze(audio_file):
|
| 83 |
if audio_file is None:
|
| 84 |
return "Nessun file caricato."
|
| 85 |
+
result = analyze_audio(audio_file)
|
| 86 |
+
ai_score = next((r["score"] for r in result if r["label"] == "AI"), 0.5)
|
| 87 |
+
verdict = "Probabile AI" if ai_score > 0.5 else "Probabile umano (autorialita' umana)"
|
|
|
|
| 88 |
return (
|
| 89 |
+
verdict + "\n\n"
|
| 90 |
+
"AI Score: " + str(round(ai_score * 100, 1)) + "%\n"
|
| 91 |
+
"Human Score: " + str(round((1 - ai_score) * 100, 1)) + "%\n\n"
|
| 92 |
+
"Soglia ADMC: 50% (configurabile nel plugin WordPress)"
|
| 93 |
)
|
| 94 |
|
| 95 |
+
|
| 96 |
+
# Gradio 4+ usa flagging_mode="never" invece di allow_flagging="never"
|
| 97 |
demo = gr.Interface(
|
| 98 |
fn=gradio_analyze,
|
| 99 |
inputs=gr.Audio(type="filepath", label="Carica brano musicale (MP3/WAV/FLAC)"),
|
| 100 |
outputs=gr.Textbox(label="Risultato analisi ADMC"),
|
| 101 |
+
title="ADMC - Rilevamento Autorialita AI nella Musica",
|
| 102 |
description=(
|
| 103 |
+
"Analizza un brano musicale per rilevare se e stato generato da AI o creato da un umano. "
|
| 104 |
+
"Parte del sistema di certificazione ADMC - Artigiani della Musica."
|
| 105 |
),
|
| 106 |
+
flagging_mode="never",
|
|
|
|
| 107 |
)
|
| 108 |
|
| 109 |
+
# Monta Gradio su FastAPI
|
|
|
|
| 110 |
app = gr.mount_gradio_app(FastAPI(), demo, path="/")
|
| 111 |
|
| 112 |
+
|
| 113 |
@app.post("/analyze")
|
| 114 |
async def api_analyze(request: Request):
|
| 115 |
"""
|
| 116 |
POST /analyze
|
| 117 |
+
Body: raw audio bytes
|
| 118 |
+
Content-Type: audio/mpeg | audio/wav | audio/flac | audio/ogg
|
| 119 |
Returns: [{"label": "AI", "score": 0.87}, {"label": "Human", "score": 0.13}]
|
| 120 |
"""
|
| 121 |
content_type = request.headers.get("content-type", "")
|
|
|
|
| 127 |
"audio/ogg": ".ogg",
|
| 128 |
"audio/aiff": ".aiff",
|
| 129 |
}
|
| 130 |
+
ext = ext_map.get(content_type.split(";")[0].strip(), ".mp3")
|
| 131 |
+
audio_data = await request.body()
|
| 132 |
|
| 133 |
+
if len(audio_data) == 0:
|
|
|
|
| 134 |
raise HTTPException(status_code=400, detail="Nessun file audio ricevuto.")
|
| 135 |
+
if len(audio_data) > 100 * 1024 * 1024:
|
| 136 |
raise HTTPException(status_code=413, detail="File troppo grande (max 100 MB).")
|
| 137 |
|
|
|
|
| 138 |
with tempfile.NamedTemporaryFile(suffix=ext, delete=False) as tmp:
|
| 139 |
+
tmp.write(audio_data)
|
| 140 |
tmp_path = tmp.name
|
| 141 |
|
| 142 |
try:
|
| 143 |
result = analyze_audio(tmp_path)
|
| 144 |
except Exception as e:
|
| 145 |
+
raise HTTPException(status_code=500, detail="Errore analisi: " + str(e))
|
| 146 |
finally:
|
| 147 |
+
if os.path.exists(tmp_path):
|
| 148 |
+
os.unlink(tmp_path)
|
| 149 |
|
| 150 |
return JSONResponse(content=result)
|
| 151 |
|
| 152 |
|
| 153 |
@app.get("/health")
|
| 154 |
async def health():
|
| 155 |
+
return {
|
| 156 |
+
"status": "ok",
|
| 157 |
+
"model": MODEL_ID,
|
| 158 |
+
"loaded": classifier is not None,
|
| 159 |
+
"gpu": torch.cuda.is_available(),
|
| 160 |
+
}
|