temporarily change to llama 8b to allow more scale

backend/tasks/create_bench_config_file.py

@@ -121,7 +121,8 @@ class CreateBenchConfigTask:
         
         # Define required models
         required_models = [
-            "Qwen/Qwen2.5-72B-Instruct"
+            # "Qwen/Qwen2.5-72B-Instruct"
+            "meta-llama/Llama-3.1-8B-Instruct"
         ]
         
         # Track found models
@@ -164,11 +165,11 @@ class CreateBenchConfigTask:
             "model_list": model_list,
             
             "model_roles": {
-                "ingestion": ["Qwen/Qwen2.5-72B-Instruct"],
-                "summarization": ["Qwen/Qwen2.5-72B-Instruct"],
+                "ingestion": ["meta-llama/Llama-3.1-8B-Instruct"],
+                "summarization": ["meta-llama/Llama-3.1-8B-Instruct"],
                 "chunking": ["intfloat/multilingual-e5-large-instruct"],
-                "single_shot_question_generation": ["Qwen/Qwen2.5-72B-Instruct"],
-                "multi_hop_question_generation": ["Qwen/Qwen2.5-72B-Instruct"],
+                "single_shot_question_generation": ["meta-llama/Llama-3.1-8B-Instruct"],
+                "multi_hop_question_generation": ["meta-llama/Llama-3.1-8B-Instruct"],
             },
             "pipeline": {
                 "ingestion": {