maintainers-highlight-submission (#877)

- Set no limits for curated authors at submit (5ac54dcdb5c0716cd809c46a38b1f975cf379412)

src/display/utils.py

@@ -1,14 +1,19 @@
 from dataclasses import dataclass, make_dataclass
+from datasets import load_dataset
 from enum import Enum
 import json
 import logging
 from datetime import datetime
 import pandas as pd
 
+from src.envs import MAINTAINERS_HIGHLIGHT_REPO
+
 # Configure logging
 logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
 
-        
+dataset = load_dataset(MAINTAINERS_HIGHLIGHT_REPO)
+curated_authors = dataset["train"][0]["CURATED_SET"]
+
 # Convert ISO 8601 dates to datetime objects for comparison
 def parse_iso8601_datetime(date_str):
     if date_str.endswith('Z'):

src/envs.py

@@ -8,6 +8,7 @@ REPO_ID = "open-llm-leaderboard/open_llm_leaderboard"
 QUEUE_REPO = "open-llm-leaderboard/requests"
 AGGREGATED_REPO = "open-llm-leaderboard/contents"
 VOTES_REPO = "open-llm-leaderboard/votes"
+MAINTAINERS_HIGHLIGHT_REPO = "open-llm-leaderboard/maintainers-highlight"
 
 HF_HOME = os.getenv("HF_HOME", ".")
 

src/submission/check_validity.py

@@ -9,7 +9,7 @@ from huggingface_hub import ModelCard
 from huggingface_hub.hf_api import ModelInfo, get_safetensors_metadata
 from transformers import AutoConfig, AutoTokenizer
 
-from src.display.utils import parse_iso8601_datetime
+from src.display.utils import parse_iso8601_datetime, curated_authors
 from src.envs import HAS_HIGHER_RATE_LIMIT
 
 
@@ -103,6 +103,10 @@ def get_model_arch(model_info: ModelInfo):
 
 
 def user_submission_permission(org_or_user, users_to_submission_dates, rate_limit_period, rate_limit_quota):
+    # No limit for curated authors
+    if org_or_user in curated_authors:
+        return True, ""
+    
     # Increase quota first if user has higher limits
     if org_or_user in HAS_HIGHER_RATE_LIMIT:
         rate_limit_quota *= 2