Update app.py

app.py

@@ -16,7 +16,7 @@ token = os.getenv("HF_TOKEN")
 device = 0 if torch.cuda.is_available() else -1
 Najeb_pipeline = pipeline("text-generation", model="sohiebwedyan/NAJEB_BOT", token=token, device=device)
 gpt2_pipeline = pipeline("text-generation", model="Qwen/Qwen-1_8B-Chat", device=device, trust_remote_code=True)
-llama2_pipeline = pipeline("text-generation", model="Harikrishnan46624/finetuned_llama2-1.1b-chat", device=device)
+#llama2_pipeline = pipeline("text-generation", model="Harikrishnan46624/finetuned_llama2-1.1b-chat", device=device)
 summarization_pipeline = pipeline("summarization", model="Falconsai/text_summarization", device=device)
 
 previous_questions = []
@@ -46,7 +46,8 @@ async def generate_Najeb(question, max_length, num_beams, temperature):
         top_p=0.85,
         temperature=temperature
     )[0]['generated_text']
-
+    
+'''
 # توليد الردود باستخدام LLaMA 2
 async def generate_llama2(question, max_length, num_beams, temperature):
     return llama2_pipeline(
@@ -58,7 +59,7 @@ async def generate_llama2(question, max_length, num_beams, temperature):
         top_k=50,
         top_p=0.95,
         temperature=temperature
-    )[0]['generated_text']
+    )[0]['generated_text']'''
 
 # التعامل مع الردود بشكل غير متزامن
 async def generate_responses_async(question, max_length=128, num_beams=2, temperature=0.5):
@@ -67,19 +68,19 @@ async def generate_responses_async(question, max_length=128, num_beams=2, temper
     # إنشاء المهام بشكل غير متزامن لتوليد الردود من الموديلات المختلفة
     gpt2_task = asyncio.create_task(generate_gpt2(question, max_length, num_beams, temperature))
     Najeb_task = asyncio.create_task(generate_Najeb(question, max_length, num_beams, temperature))
-    llama2_task = asyncio.create_task(generate_llama2(question, max_length, num_beams, temperature))
+    #llama2_task = asyncio.create_task(generate_llama2(question, max_length, num_beams, temperature))
 
     # تجميع الردود من جميع الموديلات
-    gpt2_response, Najeb_response, llama2_response = await asyncio.gather(gpt2_task, Najeb_task, llama2_task)
+    gpt2_response, Najeb_response = await asyncio.gather(gpt2_task, Najeb_task, llama2_task)
 
     # دمج الردود و تلخيصها
-    combined_responses = f"GPT-2: {gpt2_response}\nNajeb: {Najeb_response}\nLLaMA 2: {llama2_response}"
+    combined_responses = f"GPT-2: {gpt2_response}\nNajeb: {Najeb_response}"
     summarized_response = summarization_pipeline(combined_responses, max_length=150, min_length=50, do_sample=False)[0]['summary_text']
 
     return {
         "GPT-2 Answer": gpt2_response,
         "Najeb Answer": Najeb_response,
-        "LLaMA 2 Answer": llama2_response,
+        #"LLaMA 2 Answer": llama2_response,
         "Summarized Answer": summarized_response,
         "Previous Questions": "\n".join(previous_questions[-5:])
     }
@@ -91,7 +92,7 @@ def handle_mode_selection(mode, input_text, max_length, num_beams, temperature):
         return (
             f"**GPT-2 Model Response:**\n{result['GPT-2 Answer']}",
             f"**Najeb Model Response:**\n{result['Najeb Answer']}",
-            f"**LLaMA 2 Model Response:**\n{result['LLaMA 2 Answer']}",
+            #f"**LLaMA 2 Model Response:**\n{result['LLaMA 2 Answer']}",
             f"**Summarized Response:**\n{result['Summarized Answer']}",
             f"**Previous Questions:**\n{result['Previous Questions']}"
         )
@@ -245,7 +246,7 @@ gr.Interface(
     outputs=[
         gr.Markdown(label="GPT-2 Answer"),
         gr.Markdown(label="Najeb Answer"),
-        gr.Markdown(label="LLaMA 2 Answer"),
+        #gr.Markdown(label="LLaMA 2 Answer"),
         gr.Markdown(label="Summarized Answer"),
         gr.Markdown(label="Previous Questions")
     ],