a-t aya-8b

Author: DataBoySu

.github/workflows/translate.yml

@@ -1,4 +1,4 @@
-name: AI Readme Translator (Multilingual)
+name: Readme Translator (Multilingual)
 
 on:
   workflow_dispatch:
@@ -15,7 +15,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        lang: [de, fr, es, ja, zh] # German, French, Spanish, Japanese, Chinese
+        lang: [de, fr, es, ja, zh, pt, ko, hi] # German, French, Spanish, Japanese, Chinese, Portuguese, Korean, Hindi
     steps:
       - name: Checkout Code
         uses: actions/checkout@v4

README.md

@@ -4,7 +4,10 @@
   <a href="locales/README.fr.md">🇫🇷 Français</a> |
   <a href="locales/README.es.md">🇪🇸 Español</a> |
   <a href="locales/README.ja.md">🇯🇵 日本語</a> |
-  <a href="locales/README.zh.md">🇨🇳 中文</a>
+  <a href="locales/README.zh.md">🇨🇳 中文</a> |
+  <a href="locales/README.pt.md">🇵🇹 Português</a> |
+  <a href="locales/README.ko.md">🇰🇷 한국어</a> |
+  <a href="locales/README.hi.md">🇮🇳 Hindi</a>
 </div>
 
 <div style="text-align:center; margin:18px 0;">

locales/translated_readme.md

@@ -9,8 +9,11 @@
     "fr": "French",
     "es": "Spanish",
     "ja": "Japanese",
-    "zh": "Chinese (Simplified)",
-    "ru": "Russian"
+    "zh": "Simplified Chinese",
+    "ru": "Russian",
+    "pt": "Portuguese",
+    "ko": "Korean",
+    "hi": "Hindi"
 }
 
 parser = argparse.ArgumentParser()
@@ -41,6 +44,7 @@
 Perform a strict technical translation of the provided README into professional developer-level {target_lang_name}. 
 Maintain a formal tone and preserve technical terminology (e.g., GPU, CLI, vCPU).
 Keep all Markdown/HTML syntax exactly as is. 
+IMPORTANT: Do not strip or modify the top-level HTML tags (like <div> or <img>) at the beginning of the file.
 ONLY output the translated {target_lang_name} text. No talk, just translation.
 Do not add new information, do not summarize, and do not include any conversational filler or "thinking" process.<|END_OF_TURN_TOKEN|>
 <|START_OF_TURN_TOKEN|><|USER_TOKEN|>
@@ -58,9 +62,6 @@
 
 translated_content = response['choices'][0]['text'].strip()
 
-# # 0. SAFETY: Remove HTML comments if the LLM wrapped the output in them
-# translated_content = re.sub(r'^<!--\s*|(?:\s*)?-->$', '', translated_content).strip()
-
 # 1. CLEANUP: Remove markdown code fences if the LLM included them
 if translated_content.startswith("```"):
     lines = translated_content.splitlines()