Add ComfyUI models documentation and download script

- Complete MODELS.md with detailed model descriptions
- Interactive download script for 5+ specialized SDXL models
- Updated README with model selection guide
- Fix docker-compose volumes path (/root/ComfyUI)
- API workflows for OpenWebUI integration

Models documented:
- RealVisXL (portraits)
- Juggernaut XL (versatile)
- DreamShaper (art/fantasy)
- SDXL Lightning (fast)
- Proteus (realism)

docker-compose-speach.yml

@@ -0,0 +1,45 @@
+# Configuration STT (speaches) pour Fedora
+
+services:
+  speaches:
+    image: ghcr.io/speaches-ai/speaches:latest
+    container_name: speaches-stt
+    restart: unless-stopped
+    user: "0:0"
+    ports:
+      - "8200:8000"
+    environment:
+      - WHISPER__MODEL=deepdml/faster-whisper-large-v3-turbo-ct2
+      - WHISPER__LANGUAGE=fr
+      - WHISPER__COMPUTE_TYPE=int8
+      - WHISPER__DEVICE=cpu
+      - WHISPER__CPU_THREADS=4
+    volumes:
+      - speaches-models:/home/user/.cache/huggingface:Z
+    # Si ROCm fonctionne un jour, ajouter:
+    # devices:
+    #   - /dev/kfd
+    #   - /dev/dri
+    # group_add:
+    #   - "44"
+    #   - "991"
+    # environment:
+    #   - HSA_OVERRIDE_GFX_VERSION=11.0.0
+
+  speaches-init:
+    image: curlimages/curl:latest
+    container_name: speaches-init
+    restart: "no"
+    depends_on:
+      - speaches
+    entrypoint: >
+      sh -c "echo 'Attente du démarrage de speaches...' &&
+      sleep 15 &&
+      curl -sf -X POST http://speaches:8000/v1/models \
+        -H 'Content-Type: application/json' \
+        -d '{\"model\":\"deepdml/faster-whisper-large-v3-turbo-ct2\"}' &&
+      echo ' ✅ Modèle enregistré avec succès' ||
+      echo ' ⚠️ Erreur enregistrement modèle - vérifier manuellement'"
+
+volumes:
+  speaches-models: