import subprocess import gradio as gr import pandas as pd from apscheduler.schedulers.background import BackgroundScheduler 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, NUMERIC_INTERVALS, 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 from PIL import Image from dummydatagen import dummy_data_for_plot, create_metric_plot_obj_1, dummydf import copy def restart_space(): API.restart_space(repo_id=REPO_ID) # try: # print(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: # print(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() # raw_data, original_df = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS, BENCHMARK_COLS) # leaderboard_df = original_df.copy() # ( # finished_eval_queue_df, # running_eval_queue_df, # pending_eval_queue_df, # ) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS) raw_data = dummydf() methods = list(set(raw_data['Method'])) metrics = ["Chruch","Parachute","Tench","Garbage Turch","Van Gogh","Violence","Illegal Activity","Nudity"] def update_table( hidden_df: pd.DataFrame, columns_1: list, columns_2: list, columns_3: list, model1: list, ): filtered_df = select_columns(hidden_df, columns_1, columns_2, columns_3) filtered_df = filter_model1(filtered_df, model1) return filtered_df def select_columns(df: pd.DataFrame, columns_1: list, columns_2: list, columns_3: list) -> pd.DataFrame: always_here_cols = ["Method"] # We use COLS to maintain sorting all_columns = metrics if (len(columns_1)+len(columns_2) + len(columns_3)) == 0: filtered_df = df[ always_here_cols + [c for c in all_columns if c in df.columns] ] else: filtered_df = df[ always_here_cols + [c for c in all_columns if c in df.columns and (c in columns_1 or c in columns_2 or c in columns_3 ) ] ] return filtered_df def filter_model1(df: pd.DataFrame, model_query: list) -> pd.DataFrame: # Show all models if len(model_query) == 0: return df filtered_df = df filtered_df = filtered_df[filtered_df["Method"].isin(model_query)] return filtered_df 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("UnlearnDiffAtk Benchmark", elem_id="UnlearnDiffAtk-benchmark-tab-table", id=0): with gr.Row(): with gr.Column(): with gr.Row(): search_bar = gr.Textbox( placeholder=" 🔍 Search for your model (separate multiple queries with `;`) and press ENTER...", show_label=False, elem_id="search-bar", ) # with gr.Column(min_width=320): with gr.Row(): shown_columns_1 = gr.CheckboxGroup( choices=["Church","Parachute","Tench", "Garbage Truck"], label="Undersirable Objects", elem_id="column-object", interactive=True, ) with gr.Row(): shown_columns_2 = gr.CheckboxGroup( choices=["Van Gogh"], label="Undersirable Styles", elem_id="column-style", interactive=True, ) with gr.Row(): shown_columns_3 = gr.CheckboxGroup( choices=["Violence","Illegal Activity","Nudity"], label="Undersirable Concepts (Outputs that may be offensive in nature)", elem_id="column-select", interactive=True, ) with gr.Row(): shown_columns_4 = gr.Slider( 1, 100, value=40, step=1, label="Attacking Steps", info="Choose between 1 and 100", interactive=True,) leaderboard_table = gr.components.Dataframe( value= raw_data, elem_id="leaderboard-table", interactive=False, visible=True, ) game_bench_df_for_search = gr.components.Dataframe( value= raw_data, elem_id="leaderboard-table", interactive=False, visible=False, ) for selector in [shown_columns_1,shown_columns_2, shown_columns_3, model1_column]: selector.change( update_table, [ game_bench_df_for_search, shown_columns_1, shown_columns_2, shown_columns_3, model1_column, ], leaderboard_table, queue=True, ) # with gr.Row(): # shown_columns = gr.CheckboxGroup( # choices=[ # c.name # for c in fields(AutoEvalColumn) # if not c.hidden and not c.never_hidden # ], # value=[ # c.name # for c in fields(AutoEvalColumn) # if c.displayed_by_default and not c.hidden and not c.never_hidden # ], # label="Select columns to show", # elem_id="column-select", # interactive=True, # ) # with gr.Row(): # deleted_models_visibility = gr.Checkbox( # value=False, label="Show gated/private/deleted models", interactive=True # ) # with gr.Column(min_width=320): # #with gr.Box(elem_id="box-filter"): # filter_columns_type = gr.CheckboxGroup( # label="Unlearning types", # choices=[t.to_str() for t in ModelType], # value=[t.to_str() for t in ModelType], # interactive=True, # elem_id="filter-columns-type", # ) # filter_columns_precision = gr.CheckboxGroup( # label="Precision", # choices=[i.value.name for i in Precision], # value=[i.value.name for i in Precision], # interactive=True, # elem_id="filter-columns-precision", # ) # filter_columns_size = gr.CheckboxGroup( # label="Model sizes (in billions of parameters)", # choices=list(NUMERIC_INTERVALS.keys()), # value=list(NUMERIC_INTERVALS.keys()), # interactive=True, # elem_id="filter-columns-size", # ) # leaderboard_table = gr.components.Dataframe( # value=leaderboard_df[ # [c.name for c in fields(AutoEvalColumn) if c.never_hidden] # + shown_columns.value # ], # headers=[c.name for c in fields(AutoEvalColumn) if c.never_hidden] + shown_columns.value, # datatype=TYPES, # elem_id="leaderboard-table", # interactive=False, # visible=True, # ) # # Dummy leaderboard for handling the case when the user uses backspace key # hidden_leaderboard_table_for_search = gr.components.Dataframe( # value=original_df[COLS], # headers=COLS, # datatype=TYPES, # visible=False, # ) # search_bar.submit( # update_table, # [ # hidden_leaderboard_table_for_search, # shown_columns, # filter_columns_type, # filter_columns_precision, # filter_columns_size, # deleted_models_visibility, # search_bar, # ], # leaderboard_table, # ) # for selector in [shown_columns, filter_columns_type, filter_columns_precision, filter_columns_size, deleted_models_visibility]: # selector.change( # update_table, # [ # hidden_leaderboard_table_for_search, # shown_columns, # filter_columns_type, # filter_columns_precision, # filter_columns_size, # deleted_models_visibility, # search_bar, # ], # leaderboard_table, # queue=True, # ) # 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 ({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(): # 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=True): citation_button = gr.Textbox( value=CITATION_BUTTON_TEXT, label=CITATION_BUTTON_LABEL, lines=10, elem_id="citation-button", show_copy_button=True, ) scheduler = BackgroundScheduler() scheduler.add_job(restart_space, "interval", seconds=1800) scheduler.start() demo.queue().launch(share=True)