output summary unquely in zh-TW

summarize_transcript.py

@@ -18,10 +18,11 @@ def load_model():
         seed=1337,
         n_ctx=32768,       # Context size
         verbose=False,     # Reduced verbosity for cleaner output
-        n_batch=1024,
-        n_ubatch=512,
+        n_batch=2048,
+        n_ubatch=1024,
         v_type=2,
-        k_type=2
+        k_type=2,
+        n_threads=6
     )
 
     return llm
@@ -32,30 +33,23 @@ def read_transcript(file_path):
         content = f.read()
     return content
 
-def stream_summarize_transcript(llm, transcript, language='zh-TW'):
+def stream_summarize_transcript(llm, transcript):
     """
     Perform live streaming summary by getting real-time token output from the model.
 
     Args:
         llm: The loaded language model
         transcript: The full transcript to summarize
-        language: Language for the summary ('en' or 'zh-TW')
     """
 
     # Use the model's chat format based on its template
-    if language == 'en':
-        messages = [
-            {"role": "system", "content": "You are a helpful assistant that summarizes transcripts."},
-            {"role": "user", "content": f"Please summarize the following transcript:\n\n{transcript}"}
-        ]
-    else:  # Default to zh-TW
-        messages = [
-            {"role": "system", "content": "你是一個有助的助手，負責總結轉錄內容。"},
-            {"role": "user", "content": f"請總結以下內容：\n\n{transcript}"}
-        ]
+    messages = [
+        {"role": "system", "content": "你是一個有助的助手，負責總結轉錄內容。"},
+        {"role": "user", "content": f"請總結以下內容：\n\n{transcript}"}
+    ]
 
     # Generate the summary using streaming completion
-    print(f"\nStreaming {language} summary:")
+    print(f"\nStreaming zh-TW summary:")
     print("="*50)
 
     full_response = ""
@@ -102,18 +96,12 @@ def main():
     print(transcript[:500] + "..." if len(transcript) > 500 else transcript)
 
     # Summarize in Chinese (zh-TW) with streaming
-    chinese_summary = stream_summarize_transcript(llm, transcript, language='zh-TW')
-
-    # Summarize in English with streaming
-    english_summary = stream_summarize_transcript(llm, transcript, language='en')
+    chinese_summary = stream_summarize_transcript(llm, transcript)
 
     # Save summaries to files
     with open("/home/luigi/tiny-scribe/chinese_summary.txt", 'w', encoding='utf-8') as f:
         f.write(chinese_summary)
 
-    with open("/home/luigi/tiny-scribe/english_summary.txt", 'w', encoding='utf-8') as f:
-        f.write(english_summary)
-
     print("\nSummaries saved to chinese_summary.txt and english_summary.txt.")
 
     # Clean up