Spaces:
Sleeping
Sleeping
Fix threading lock
Browse files- api_server.py +30 -61
api_server.py
CHANGED
|
@@ -1,10 +1,6 @@
|
|
| 1 |
#!/usr/bin/env python3
|
| 2 |
-
"""
|
| 3 |
-
API Flask pour générer des images
|
| 4 |
-
Endpoint unique: /generate
|
| 5 |
-
"""
|
| 6 |
-
|
| 7 |
import torch
|
|
|
|
| 8 |
from pathlib import Path
|
| 9 |
from flask import Flask, request, jsonify, send_file
|
| 10 |
from diffusers import StableDiffusionPipeline
|
|
@@ -13,41 +9,26 @@ import io
|
|
| 13 |
|
| 14 |
app = Flask(__name__)
|
| 15 |
|
| 16 |
-
# Configuration
|
| 17 |
OUTPUT_DIR = Path("/app/generated_images")
|
| 18 |
OUTPUT_DIR.mkdir(exist_ok=True, parents=True)
|
| 19 |
-
|
| 20 |
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
|
| 21 |
-
|
| 22 |
-
# Variable globale
|
| 23 |
pipeline = None
|
| 24 |
model_loaded = False
|
|
|
|
| 25 |
|
| 26 |
def load_model():
|
| 27 |
-
"""Charge le modèle au démarrage"""
|
| 28 |
global pipeline, model_loaded
|
| 29 |
-
|
| 30 |
-
print("\n" + "="*70)
|
| 31 |
-
print("🤖 Chargement du modèle fusionné...")
|
| 32 |
-
print("="*70 + "\n")
|
| 33 |
-
|
| 34 |
try:
|
| 35 |
print(f"📱 Appareil: {DEVICE}")
|
| 36 |
-
|
| 37 |
dtype = torch.float32 if DEVICE == "cpu" else torch.float16
|
| 38 |
-
|
| 39 |
pipeline = StableDiffusionPipeline.from_pretrained(
|
| 40 |
"K2MAR/mon-modele-sd",
|
| 41 |
torch_dtype=dtype,
|
| 42 |
safety_checker=None
|
| 43 |
).to(DEVICE)
|
| 44 |
-
|
| 45 |
pipeline.enable_attention_slicing()
|
| 46 |
-
|
| 47 |
model_loaded = True
|
| 48 |
-
print("="*70)
|
| 49 |
print("✅ Modèle prêt!")
|
| 50 |
-
print("="*70 + "\n")
|
| 51 |
return True
|
| 52 |
except Exception as e:
|
| 53 |
print(f"❌ Erreur: {e}")
|
|
@@ -58,71 +39,59 @@ def health():
|
|
| 58 |
return jsonify({
|
| 59 |
"status": "ok" if model_loaded else "loading",
|
| 60 |
"device": DEVICE,
|
| 61 |
-
"model_loaded": model_loaded
|
|
|
|
| 62 |
})
|
| 63 |
|
| 64 |
@app.route('/generate', methods=['POST'])
|
| 65 |
def generate():
|
| 66 |
if not model_loaded:
|
| 67 |
return jsonify({"error": "Model not loaded"}), 503
|
| 68 |
-
|
|
|
|
|
|
|
|
|
|
| 69 |
try:
|
| 70 |
data = request.get_json()
|
| 71 |
-
|
| 72 |
if not data or "prompt" not in data:
|
| 73 |
-
return jsonify({"error": "Missing 'prompt'
|
| 74 |
-
|
| 75 |
prompt = data.get("prompt", "")
|
| 76 |
-
steps = int(data.get("steps",
|
| 77 |
guidance_scale = float(data.get("guidance_scale", 7.5))
|
| 78 |
-
|
| 79 |
if not prompt:
|
| 80 |
return jsonify({"error": "Prompt cannot be empty"}), 400
|
| 81 |
-
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
|
| 93 |
-
|
| 94 |
-
).images[0]
|
| 95 |
-
|
| 96 |
-
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
|
| 97 |
-
filepath = OUTPUT_DIR / f"generated_{timestamp}.png"
|
| 98 |
-
image.save(filepath)
|
| 99 |
-
|
| 100 |
img_io = io.BytesIO()
|
| 101 |
image.save(img_io, 'PNG')
|
| 102 |
img_io.seek(0)
|
| 103 |
-
|
|
|
|
| 104 |
return send_file(img_io, mimetype='image/png')
|
| 105 |
-
|
| 106 |
except Exception as e:
|
| 107 |
print(f"❌ Erreur: {str(e)}\n")
|
| 108 |
return jsonify({"error": str(e)}), 500
|
| 109 |
|
| 110 |
@app.route('/', methods=['GET'])
|
| 111 |
def home():
|
| 112 |
-
return jsonify({
|
| 113 |
-
"service": "LoRA Solar Panel Generator API",
|
| 114 |
-
"version": "1.0",
|
| 115 |
-
"device": DEVICE,
|
| 116 |
-
"model_loaded": model_loaded,
|
| 117 |
-
"endpoints": {
|
| 118 |
-
"health": "GET /health",
|
| 119 |
-
"generate": "POST /generate"
|
| 120 |
-
}
|
| 121 |
-
})
|
| 122 |
|
| 123 |
if __name__ == '__main__':
|
| 124 |
if not load_model():
|
| 125 |
exit(1)
|
| 126 |
-
|
| 127 |
-
print("\n🚀 Serveur démarrage sur 0.0.0.0:7860\n")
|
| 128 |
app.run(host='0.0.0.0', port=7860, debug=False, threaded=True)
|
|
|
|
| 1 |
#!/usr/bin/env python3
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
import torch
|
| 3 |
+
import threading
|
| 4 |
from pathlib import Path
|
| 5 |
from flask import Flask, request, jsonify, send_file
|
| 6 |
from diffusers import StableDiffusionPipeline
|
|
|
|
| 9 |
|
| 10 |
app = Flask(__name__)
|
| 11 |
|
|
|
|
| 12 |
OUTPUT_DIR = Path("/app/generated_images")
|
| 13 |
OUTPUT_DIR.mkdir(exist_ok=True, parents=True)
|
|
|
|
| 14 |
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
|
|
|
|
|
|
|
| 15 |
pipeline = None
|
| 16 |
model_loaded = False
|
| 17 |
+
lock = threading.Lock()
|
| 18 |
|
| 19 |
def load_model():
|
|
|
|
| 20 |
global pipeline, model_loaded
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
try:
|
| 22 |
print(f"📱 Appareil: {DEVICE}")
|
|
|
|
| 23 |
dtype = torch.float32 if DEVICE == "cpu" else torch.float16
|
|
|
|
| 24 |
pipeline = StableDiffusionPipeline.from_pretrained(
|
| 25 |
"K2MAR/mon-modele-sd",
|
| 26 |
torch_dtype=dtype,
|
| 27 |
safety_checker=None
|
| 28 |
).to(DEVICE)
|
|
|
|
| 29 |
pipeline.enable_attention_slicing()
|
|
|
|
| 30 |
model_loaded = True
|
|
|
|
| 31 |
print("✅ Modèle prêt!")
|
|
|
|
| 32 |
return True
|
| 33 |
except Exception as e:
|
| 34 |
print(f"❌ Erreur: {e}")
|
|
|
|
| 39 |
return jsonify({
|
| 40 |
"status": "ok" if model_loaded else "loading",
|
| 41 |
"device": DEVICE,
|
| 42 |
+
"model_loaded": model_loaded,
|
| 43 |
+
"busy": lock.locked()
|
| 44 |
})
|
| 45 |
|
| 46 |
@app.route('/generate', methods=['POST'])
|
| 47 |
def generate():
|
| 48 |
if not model_loaded:
|
| 49 |
return jsonify({"error": "Model not loaded"}), 503
|
| 50 |
+
|
| 51 |
+
if lock.locked():
|
| 52 |
+
return jsonify({"error": "Server busy, try again later"}), 503
|
| 53 |
+
|
| 54 |
try:
|
| 55 |
data = request.get_json()
|
|
|
|
| 56 |
if not data or "prompt" not in data:
|
| 57 |
+
return jsonify({"error": "Missing 'prompt'"}), 400
|
| 58 |
+
|
| 59 |
prompt = data.get("prompt", "")
|
| 60 |
+
steps = min(int(data.get("steps", 20)), 30)
|
| 61 |
guidance_scale = float(data.get("guidance_scale", 7.5))
|
| 62 |
+
|
| 63 |
if not prompt:
|
| 64 |
return jsonify({"error": "Prompt cannot be empty"}), 400
|
| 65 |
+
|
| 66 |
+
print(f"\n🎨 Génération: {prompt} ({steps} steps)")
|
| 67 |
+
|
| 68 |
+
with lock:
|
| 69 |
+
with torch.no_grad():
|
| 70 |
+
image = pipeline(
|
| 71 |
+
prompt,
|
| 72 |
+
num_inference_steps=steps,
|
| 73 |
+
guidance_scale=guidance_scale,
|
| 74 |
+
height=512,
|
| 75 |
+
width=512
|
| 76 |
+
).images[0]
|
| 77 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 78 |
img_io = io.BytesIO()
|
| 79 |
image.save(img_io, 'PNG')
|
| 80 |
img_io.seek(0)
|
| 81 |
+
|
| 82 |
+
print(f"✅ Image générée!\n")
|
| 83 |
return send_file(img_io, mimetype='image/png')
|
| 84 |
+
|
| 85 |
except Exception as e:
|
| 86 |
print(f"❌ Erreur: {str(e)}\n")
|
| 87 |
return jsonify({"error": str(e)}), 500
|
| 88 |
|
| 89 |
@app.route('/', methods=['GET'])
|
| 90 |
def home():
|
| 91 |
+
return jsonify({"service": "SD API", "model_loaded": model_loaded, "device": DEVICE})
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 92 |
|
| 93 |
if __name__ == '__main__':
|
| 94 |
if not load_model():
|
| 95 |
exit(1)
|
| 96 |
+
print("\n🚀 Serveur sur 0.0.0.0:7860\n")
|
|
|
|
| 97 |
app.run(host='0.0.0.0', port=7860, debug=False, threaded=True)
|