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>

config.yaml

@@ -0,0 +1,45 @@
+# Kokoro TTS Wyoming Server Configuration
+
+server:
+  uri: tcp://0.0.0.0:10200
+
+tts:
+  device: cuda          # ROCm presents as 'cuda' to PyTorch via HIP
+  language: a           # a=American English, b=British English, e=Spanish,
+                        # f=French, h=Hindi, i=Italian, j=Japanese,
+                        # p=Portuguese, z=Mandarin
+
+  default_voice: af_heart
+  default_speed: 1.0
+
+  # Voices to advertise to Home Assistant.
+  # Add or remove entries as needed.
+  # Voice files are downloaded from HuggingFace on first use.
+  voices:
+    - name: af_heart
+      description: "Heart (Female, American English)"
+      language: en-us
+
+    - name: af_bella
+      description: "Bella (Female, American English)"
+      language: en-us
+
+    - name: af_nova
+      description: "Nova (Female, American English)"
+      language: en-us
+
+    - name: am_michael
+      description: "Michael (Male, American English)"
+      language: en-us
+
+    - name: am_fenrir
+      description: "Fenrir (Male, American English)"
+      language: en-us
+
+    - name: bf_emma
+      description: "Emma (Female, British English)"
+      language: en-gb
+
+    - name: bm_george
+      description: "George (Male, British English)"
+      language: en-gb