Change provider in config.py

yourbench_space/config.py

@@ -18,12 +18,12 @@ def generate_base_config(hf_org: str, hf_dataset_name: str, session_uid: str):
         "model_list": [
             {
                 "model_name": "Qwen/Qwen2.5-VL-72B-Instruct",
-                "provider": "novita",
+                "provider": "nebius",
                 "max_concurrent_requests": 32,
             },
             {
                 "model_name": "Qwen/Qwen2.5-72B-Instruct",
-                "provider": "novita",
+                "provider": "nebius",
                 "max_concurrent_requests": 32,
             },
         ],
@@ -36,26 +36,32 @@ def generate_base_config(hf_org: str, hf_dataset_name: str, session_uid: str):
         },
         "pipeline": {
             "ingestion": {
+                "run": False,
                 "source_documents_dir": f"{PATH}/{session_uid}/uploaded_files/",
                 "output_dir": f"{PATH}/{session_uid}/ingested",
-                "run": True,
             },
             "upload_ingest_to_hub": {
-                "source_documents_dir": f"{PATH}/{session_uid}/ingested",
                 "run": True,
+                "source_documents_dir": f"{PATH}/{session_uid}/ingested",
             },
             "summarization": {
                 "run": True,
+                "max_tokens": 16384,
+                "token_overlap": 64,
+                "encoding_name": "cl100k_base",
             },
             "chunking": {
                 "run": True,
                 "chunking_configuration": {
+                    "chunking_mode": "fast_chunking",
+                    "l_max_tokens": 256,
+                    "token_overlap": 64,
+                    "encoding_name": "cl100k_base",
                     "l_min_tokens": 64,
-                    "l_max_tokens": 128,
                     "tau_threshold": 0.8,
                     "h_min": 2,
-                    "h_max": 5,
-                    "num_multihops_factor": 2,
+                    "h_max": 3,
+                    "num_multihops_factor": 3,
                 },
             },
             "single_shot_question_generation": {
@@ -64,7 +70,7 @@ def generate_base_config(hf_org: str, hf_dataset_name: str, session_uid: str):
                 "chunk_sampling": {
                     "mode": "count",
                     "value": 5,
-                    "random_seed": 123,
+                    "random_seed": 49,
                 },
             },
             "multi_hop_question_generation": {
@@ -79,6 +85,9 @@ def generate_base_config(hf_org: str, hf_dataset_name: str, session_uid: str):
             "lighteval": {
                 "run": True,
             },
+            "citation_score_filtering": {
+                "run": True,
+            },
         },
     }