import logging import gradio as gr from gradio_leaderboard import Leaderboard, ColumnFilter, SelectColumns import pandas as pd from apscheduler.schedulers.background import BackgroundScheduler from apscheduler.executors.pool import ThreadPoolExecutor from apscheduler.jobstores.memory import MemoryJobStore from huggingface_hub import snapshot_download from src.about import ( CITATION_BUTTON_LABEL, CITATION_BUTTON_TEXT, EVALUATION_QUEUE_TEXT, INTRODUCTION_TEXT, LLM_BENCHMARKS_TEXT, TITLE, ) from src.display.css_html_js import custom_css from src.display.utils import ( BENCHMARK_COLS, COLS, EVAL_COLS, EVAL_TYPES, AutoEvalColumn, ModelType, fields, WeightType, Precision ) from src.envs import API, EVAL_REQUESTS_PATH, EVAL_RESULTS_PATH, QUEUE_REPO, REPO_ID, RESULTS_REPO, TOKEN from src.populate import get_evaluation_queue_df, get_leaderboard_df from src.submission.submit import add_new_eval # Configure Logging logging.basicConfig(level=logging.INFO) logger = logging.getLogger(__name__) # Initialize Scheduler scheduler = BackgroundScheduler( jobstores={'default': MemoryJobStore()}, executors={'default': ThreadPoolExecutor(10)}, job_defaults={'coalesce': False, 'max_instances': 1}, ) scheduler.start() def restart_space(): API.restart_space(repo_id=REPO_ID) ### Space initialisation try: logger.info(f"Downloading evaluation requests from {QUEUE_REPO} to {EVAL_REQUESTS_PATH}") snapshot_download( repo_id=QUEUE_REPO, local_dir=EVAL_REQUESTS_PATH, repo_type="dataset", tqdm_class=None, etag_timeout=30, token=TOKEN ) except Exception: restart_space() try: logger.info(f"Downloading evaluation results from {RESULTS_REPO} to {EVAL_RESULTS_PATH}") snapshot_download( repo_id=RESULTS_REPO, local_dir=EVAL_RESULTS_PATH, repo_type="dataset", tqdm_class=None, etag_timeout=30, token=TOKEN ) except Exception: restart_space() LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS, BENCHMARK_COLS) ( finished_eval_queue_df, running_eval_queue_df, pending_eval_queue_df, ) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS) def init_leaderboard(dataframe): if dataframe is None or dataframe.empty: raise ValueError("Leaderboard DataFrame is empty or None.") return Leaderboard( value=dataframe, datatype=[c.type for c in fields(AutoEvalColumn)], select_columns=SelectColumns( default_selection=[c.name for c in fields(AutoEvalColumn) if c.displayed_by_default], cant_deselect=[c.name for c in fields(AutoEvalColumn) if c.never_hidden], label="Select Columns to Display:", ), search_columns=[AutoEvalColumn.model.name, AutoEvalColumn.license.name], hide_columns=[c.name for c in fields(AutoEvalColumn) if c.hidden], filter_columns=[ ColumnFilter(AutoEvalColumn.model_type.name, type="checkboxgroup", label="Model types"), ColumnFilter(AutoEvalColumn.precision.name, type="checkboxgroup", label="Precision"), ColumnFilter( AutoEvalColumn.params.name, type="slider", min=0.01, max=150, label="Select the number of parameters (B)", ), ColumnFilter( AutoEvalColumn.still_on_hub.name, type="boolean", label="Deleted/incomplete", default=True ), ], bool_checkboxgroup_label="Hide models", interactive=False, ) def start_evaluation(row): logger.info(f"Starting evaluation for row ID {row.get('id')}") # Implementation to start evaluation pass def monitor_evaluation(row): logger.info(f"Monitoring evaluation for row ID {row.get('id')}") # Implementation to monitor evaluation pass def initiate_new_evaluation(row): logger.info(f"Initiating new evaluation for row ID {row.get('id')}") # Implementation to initiate new evaluation pass def finalize_evaluation(row): logger.info(f"Finalizing evaluation for row ID {row.get('id')}") # Implementation to finalize evaluation pass def process_evaluation_queue(): """Process pending evaluation requests.""" logger.info("Starting processing of evaluation queue") try: # Retrieve evaluation queues finished_eval_queue_df, running_eval_queue_df, pending_eval_queue_df = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS) # Assign statuses to each DataFrame finished_eval_queue_df['status'] = 'FINISHED' running_eval_queue_df['status'] = 'RUNNING' pending_eval_queue_df['status'] = 'PENDING' # Handle PENDING_NEW_EVAL if 'needs_new_eval' in pending_eval_queue_df.columns: pending_new_eval_df = pending_eval_queue_df[pending_eval_queue_df['needs_new_eval']].copy() pending_new_eval_df['status'] = 'PENDING_NEW_EVAL' pending_eval_queue_df = pending_eval_queue_df[~pending_eval_queue_df['needs_new_eval']] else: pending_new_eval_df = pd.DataFrame() # Combine all queues into a single DataFrame full_queue_df = pd.concat([ finished_eval_queue_df, running_eval_queue_df, pending_eval_queue_df, pending_new_eval_df ], ignore_index=True) logger.debug(f"Combined queue has {len(full_queue_df)} entries") # Process each entry based on status for _, row in full_queue_df.iterrows(): status = row['status'] logger.debug(f"Processing row ID {row.get('id')} with status {status}") if status == 'PENDING': start_evaluation(row) elif status == 'RUNNING': monitor_evaluation(row) elif status == 'PENDING_NEW_EVAL': initiate_new_evaluation(row) elif status == 'FINISHED': finalize_evaluation(row) else: logger.warning(f"Unknown status '{status}' for row ID {row.get('id')}") logger.info("Completed processing of evaluation queue") return finished_eval_queue_df, running_eval_queue_df, pending_eval_queue_df except Exception as e: logger.error(f"Error processing evaluation queue: {e}", exc_info=True) 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("🏅 LLM Benchmark", elem_id="llm-benchmark-tab-table", id=0): leaderboard = init_leaderboard(LEADERBOARD_DF) with gr.TabItem("📝 About", elem_id="llm-benchmark-tab-table", id=2): gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text") with gr.TabItem("🚀 Submit here! ", elem_id="llm-benchmark-tab-table", id=3): with gr.Column(): with gr.Row(): gr.Markdown(EVALUATION_QUEUE_TEXT, elem_classes="markdown-text") with gr.Column(): with gr.Accordion( f"✅ Finished Evaluations", 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", 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", 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, ) # Process the evaluation queue every 2 minutes timer = gr.Timer(120, active=True) timer.tick(process_evaluation_queue, inputs=[], outputs=[finished_eval_table, running_eval_table, pending_eval_table]) with gr.Row(): gr.Markdown("# ✉️✨ Submit your model here!", elem_classes="markdown-text") with gr.Row(): 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=[t.to_str(" : ") for t in ModelType if t != ModelType.Unknown], label="Model type", multiselect=False, value=None, interactive=True, ) with gr.Column(): precision = gr.Dropdown( choices=[i.value.name for i in Precision if i != Precision.Unknown], label="Precision", multiselect=False, value="float16", interactive=True, ) weight_type = gr.Dropdown( choices=[i.value.name for i in WeightType], 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): citation_button = gr.Textbox( value=CITATION_BUTTON_TEXT, label=CITATION_BUTTON_LABEL, lines=20, elem_id="citation-button", show_copy_button=True, ) demo.queue(default_concurrency_limit=40).launch()