Spaces:
Running
Running
| import gradio as gr | |
| import pandas as pd | |
| from apscheduler.schedulers.background import BackgroundScheduler | |
| # Removed Hugging Face Hub imports as they are not needed for the simplified leaderboard | |
| # from huggingface_hub import snapshot_download, HfApi | |
| from src.about import ( # Assuming these still exist and are relevant for other tabs | |
| CITATION_BUTTON_LABEL, | |
| CITATION_BUTTON_TEXT, | |
| EVALUATION_QUEUE_TEXT, | |
| INTRODUCTION_TEXT, | |
| LLM_BENCHMARKS_TEXT, | |
| TITLE, | |
| ) | |
| from src.display.css_html_js import custom_css # Keep custom CSS | |
| # Removed utils imports related to the old leaderboard | |
| # from src.display.utils import (...) | |
| from src.envs import REPO_ID # Keep if needed for restart_space or other functions | |
| # Removed constants related to old data paths and repos if not needed elsewhere | |
| # from src.envs import API, EVAL_REQUESTS_PATH, EVAL_RESULTS_PATH, QUEUE_REPO, REPO_ID, RESULTS_REPO, TOKEN | |
| # Removed old data processing functions | |
| # from src.populate import get_evaluation_queue_df, get_leaderboard_df | |
| from src.submission.submit import add_new_eval # Keep submission logic | |
| # --- Elo Leaderboard Configuration --- | |
| # Data from the table provided by the user | |
| data = [ | |
| {'model': 'gpt-4o-mini', 'MLE-Lite_Elo': 753, 'Tabular_Elo': 839, 'NLP_Elo': 758, 'CV_Elo': 754, 'Overall': 778}, | |
| {'model': 'gpt-4o', 'MLE-Lite_Elo': 830, 'Tabular_Elo': 861, 'NLP_Elo': 903, 'CV_Elo': 761, 'Overall': 841}, | |
| {'model': 'o3-mini', 'MLE-Lite_Elo': 1108, 'Tabular_Elo': 1019, 'NLP_Elo': 1056, 'CV_Elo': 1207, 'Overall': 1096}, | |
| # Renamed 'DeepSeek-v3' to match previous list - adjust if needed | |
| {'model': 'deepseek-v3', 'MLE-Lite_Elo': 1004, 'Tabular_Elo': 1015, 'NLP_Elo': 1028, 'CV_Elo': 1067, 'Overall': 1023}, | |
| # Renamed 'DeepSeek-r1' to match previous list - adjust if needed | |
| {'model': 'deepseek-r1', 'MLE-Lite_Elo': 1137, 'Tabular_Elo': 1053, 'NLP_Elo': 1103, 'CV_Elo': 1083, 'Overall': 1100}, | |
| # Renamed 'Gemini-2.0-Flash' to match previous list - adjust if needed | |
| {'model': 'gemini-2.0-flash', 'MLE-Lite_Elo': 847, 'Tabular_Elo': 923, 'NLP_Elo': 860, 'CV_Elo': 978, 'Overall': 895}, | |
| # Renamed 'Gemini-2.0-Pro' to match previous list - adjust if needed | |
| {'model': 'gemini-2.0-pro', 'MLE-Lite_Elo': 1064, 'Tabular_Elo': 1139, 'NLP_Elo': 1028, 'CV_Elo': 973, 'Overall': 1054}, | |
| # Renamed 'Gemini-2.5-Pro' to match previous list - adjust if needed | |
| {'model': 'gemini-2.5-pro', 'MLE-Lite_Elo': 1257, 'Tabular_Elo': 1150, 'NLP_Elo': 1266, 'CV_Elo': 1177, 'Overall': 1214}, | |
| ] | |
| # Create a master DataFrame | |
| master_df = pd.DataFrame(data) | |
| # Define categories for selection (user-facing) | |
| CATEGORIES = ["MLE-Lite", "Tabular", "NLP", "CV", "Overall"] | |
| DEFAULT_CATEGORY = "Overall" # Set a default category | |
| # Map user-facing categories to DataFrame column names | |
| category_to_column = { | |
| "MLE-Lite": "MLE-Lite_Elo", | |
| "Tabular": "Tabular_Elo", | |
| "NLP": "NLP_Elo", | |
| "CV": "CV_Elo", | |
| "Overall": "Overall" | |
| } | |
| # --- Helper function to update leaderboard --- | |
| def update_leaderboard(category): | |
| """ | |
| Selects the relevant columns for the category, renames the score column | |
| to 'Elo Score', sorts by score descending, and returns the DataFrame. | |
| """ | |
| score_column = category_to_column.get(category) | |
| if score_column is None or score_column not in master_df.columns: | |
| # Fallback if category or column is invalid | |
| print(f"Warning: Invalid category '{category}' or column '{score_column}'. Falling back to default.") | |
| score_column = category_to_column[DEFAULT_CATEGORY] | |
| if score_column not in master_df.columns: # Check fallback column too | |
| return pd.DataFrame({"Model": [], "Elo Score": []}) # Return empty if still invalid | |
| # Select model and the specific score column | |
| df = master_df[['model', score_column]].copy() | |
| # Rename the score column to 'Elo Score' for consistent display | |
| df.rename(columns={score_column: 'Elo Score'}, inplace=True) | |
| # Sort by 'Elo Score' descending | |
| df.sort_values(by='Elo Score', ascending=False, inplace=True) | |
| # Reset index for cleaner display (optional) | |
| df.reset_index(drop=True, inplace=True) | |
| return df | |
| # --- Mock/Placeholder functions/data for other tabs --- | |
| # (Same as previous version - providing empty data) | |
| print("Warning: Evaluation queue data fetching is disabled/mocked due to leaderboard changes.") | |
| finished_eval_queue_df = pd.DataFrame(columns=["Model", "Status", "Requested", "Started"]) | |
| running_eval_queue_df = pd.DataFrame(columns=["Model", "Status", "Requested", "Started"]) | |
| pending_eval_queue_df = pd.DataFrame(columns=["Model", "Status", "Requested", "Started"]) | |
| EVAL_COLS = ["Model", "Status", "Requested", "Started"] # Define for the dataframe headers | |
| EVAL_TYPES = ["str", "str", "str", "str"] # Define for the dataframe types | |
| # --- Keep restart function if relevant --- | |
| # (Same as previous version) | |
| def restart_space(): | |
| print(f"Attempting to restart space: {REPO_ID}") | |
| # Replace with your actual space restart mechanism if needed | |
| # --- Gradio App Definition --- | |
| demo = gr.Blocks(css=custom_css) | |
| with demo: | |
| gr.HTML(TITLE) | |
| gr.Markdown(INTRODUCTION_TEXT, elem_classes="markdown-text") | |
| with gr.Tabs(elem_classes="tab-buttons") as tabs: | |
| with gr.TabItem("🏅 MLE-Dojo Benchmark", elem_id="llm-benchmark-tab-table", id=0): | |
| with gr.Column(): | |
| gr.Markdown("## Model Elo Rankings") # New title for the section | |
| category_selector = gr.Radio( | |
| choices=CATEGORIES, | |
| label="Select Category to Sort By", # Updated label | |
| value=DEFAULT_CATEGORY, # Default selection | |
| interactive=True, | |
| container=False, | |
| ) | |
| leaderboard_df_component = gr.Dataframe( | |
| # Initialize with sorted data for the default category | |
| value=update_leaderboard(DEFAULT_CATEGORY), | |
| headers=["Model", "Elo Score"], | |
| datatype=["str", "number"], | |
| interactive=False, | |
| # Adjust row count based on the number of models | |
| row_count=(len(master_df), "fixed"), | |
| col_count=(2, "fixed"), | |
| ) | |
| # Link the radio button change to the update function | |
| category_selector.change( | |
| fn=update_leaderboard, | |
| inputs=category_selector, | |
| outputs=leaderboard_df_component | |
| ) | |
| with gr.TabItem("📝 About", elem_id="llm-benchmark-tab-table", id=2): | |
| # (Content unchanged) | |
| gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text") | |
| # with gr.TabItem("🚀 Submit here! ", elem_id="llm-benchmark-tab-table", id=3): | |
| # # (Content unchanged, still uses potentially empty/mock queue data) | |
| # with gr.Column(): | |
| # with gr.Row(): | |
| # gr.Markdown(EVALUATION_QUEUE_TEXT, elem_classes="markdown-text") | |
| # with gr.Column(): | |
| # with gr.Accordion( | |
| # f"✅ Finished Evaluations ({len(finished_eval_queue_df)})", | |
| # open=False, | |
| # ): | |
| # with gr.Row(): | |
| # finished_eval_table = gr.components.Dataframe( | |
| # value=finished_eval_queue_df, | |
| # headers=EVAL_COLS, | |
| # datatype=EVAL_TYPES, | |
| # row_count=5, | |
| # ) | |
| # with gr.Accordion( | |
| # f"🔄 Running Evaluation Queue ({len(running_eval_queue_df)})", | |
| # open=False, | |
| # ): | |
| # with gr.Row(): | |
| # running_eval_table = gr.components.Dataframe( | |
| # value=running_eval_queue_df, | |
| # headers=EVAL_COLS, | |
| # datatype=EVAL_TYPES, | |
| # row_count=5, | |
| # ) | |
| # with gr.Accordion( | |
| # f"⏳ Pending Evaluation Queue ({len(pending_eval_queue_df)})", | |
| # open=False, | |
| # ): | |
| # with gr.Row(): | |
| # pending_eval_table = gr.components.Dataframe( | |
| # value=pending_eval_queue_df, | |
| # headers=EVAL_COLS, | |
| # datatype=EVAL_TYPES, | |
| # row_count=5, | |
| # ) | |
| # with gr.Row(): | |
| # gr.Markdown("# ✉️✨ Submit your model here!", elem_classes="markdown-text") | |
| # with gr.Row(): | |
| # # Submission form - kept as is | |
| # with gr.Column(): | |
| # model_name_textbox = gr.Textbox(label="Model name") | |
| # revision_name_textbox = gr.Textbox(label="Revision commit", placeholder="main") | |
| # model_type = gr.Dropdown( | |
| # choices=["Type A", "Type B", "Type C"], # Example choices | |
| # label="Model type", | |
| # multiselect=False, | |
| # value=None, | |
| # interactive=True, | |
| # ) | |
| # with gr.Column(): | |
| # precision = gr.Dropdown( | |
| # choices=["float16", "bfloat16", "float32", "int8"], # Example choices | |
| # label="Precision", | |
| # multiselect=False, | |
| # value="float16", | |
| # interactive=True, | |
| # ) | |
| # weight_type = gr.Dropdown( | |
| # choices=["Original", "Adapter", "Delta"], # Example choices | |
| # label="Weights type", | |
| # multiselect=False, | |
| # value="Original", | |
| # interactive=True, | |
| # ) | |
| # base_model_name_textbox = gr.Textbox(label="Base model (for delta or adapter weights)") | |
| # submit_button = gr.Button("Submit Eval") | |
| # submission_result = gr.Markdown() | |
| # submit_button.click( | |
| # add_new_eval, | |
| # [ | |
| # model_name_textbox, | |
| # base_model_name_textbox, | |
| # revision_name_textbox, | |
| # precision, | |
| # weight_type, | |
| # model_type, | |
| # ], | |
| # submission_result, | |
| # ) | |
| with gr.Row(): | |
| with gr.Accordion("📙 Citation", open=False): | |
| # (Content unchanged) | |
| citation_button = gr.Textbox( | |
| value=CITATION_BUTTON_TEXT, | |
| label=CITATION_BUTTON_LABEL, | |
| lines=20, | |
| elem_id="citation-button", | |
| show_copy_button=True, | |
| ) | |
| # --- Keep scheduler if relevant --- | |
| # scheduler = BackgroundScheduler() | |
| # scheduler.add_job(restart_space, "interval", seconds=1800) # Restart every 30 mins | |
| # scheduler.start() | |
| # --- Launch the app --- | |
| demo.launch() |