Replace upstream library with ROCm/Wyoming deployment project

Remove original Kokoro library source, demo, examples, tests, JS port,
and GitHub config. Add Dockerfile (ROCm 6.1 / PyTorch 2.5.1), Wyoming
TCP server, docker-compose with GPU passthrough, config, entrypoint,
and Gitea Actions build workflow.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

docker-compose.yml

@@ -0,0 +1,45 @@
+services:
+  kokoro-tts:
+    build: .
+    image: kokoro-rocm-wyoming:latest
+    container_name: kokoro-tts
+    restart: unless-stopped
+
+    # ROCm GPU passthrough
+    devices:
+      - /dev/kfd:/dev/kfd
+      - /dev/dri:/dev/dri
+    group_add:
+      - video
+      - render
+
+    ports:
+      - "10200:10200"
+
+    volumes:
+      # Persist HuggingFace model/voice cache so downloads survive container restarts
+      - hf_cache:/root/.cache/huggingface
+
+    environment:
+      - HIP_VISIBLE_DEVICES=0
+      # Uncomment if ROCm does not recognise the 6700 XT automatically:
+      # - HSA_OVERRIDE_GFX_VERSION=10.3.0
+
+    healthcheck:
+      test:
+        - CMD
+        - python3
+        - -c
+        - |
+          import socket
+          s = socket.socket()
+          s.settimeout(5)
+          s.connect(('localhost', 10200))
+          s.close()
+      interval: 30s
+      timeout: 10s
+      retries: 3
+      start_period: 60s
+
+volumes:
+  hf_cache: