leaderboard / app.py
Muennighoff's picture
Updates
003d24d
raw
history blame
15.2 kB
import gradio as gr
import pandas as pd
from huggingface_hub import HfApi, hf_hub_download
from huggingface_hub.repocard import metadata_load
path = f"https://huggingface.co/api/spaces"
TASKS = [
"BitextMining",
"Classification",
"Clustering",
"PairClassification",
"Reranking",
"Retrieval",
"STS",
"Summarization",
]
TASK_LIST_CLASSIFICATION = [
"AmazonCounterfactualClassification (en)",
"AmazonPolarityClassification",
"AmazonReviewsClassification (en)",
"Banking77Classification",
"EmotionClassification",
"ImdbClassification",
"MassiveIntentClassification (en)",
"MassiveScenarioClassification (en)",
"MTOPDomainClassification (en)",
"MTOPIntentClassification (en)",
"ToxicConversationsClassification",
"TweetSentimentExtractionClassification",
]
TASK_LIST_CLUSTERING = [
"ArxivClusteringP2P",
"ArxivClusteringS2S",
"BiorxivClusteringP2P",
"BiorxivClusteringS2S",
"MedrxivClusteringP2P",
"MedrxivClusteringS2S",
"RedditClustering",
"RedditClusteringP2P",
"StackExchangeClustering",
"StackExchangeClusteringP2P",
"TwentyNewsgroupsClustering",
]
TASK_LIST_PAIR_CLASSIFICATION = [
"SprintDuplicateQuestions",
"TwitterSemEval2015",
"TwitterURLCorpus",
]
TASK_LIST_RERANKING = [
"AskUbuntuDupQuestions",
"MindSmallReranking",
"SciDocsRR",
"StackOverflowDupQuestions",
]
TASK_LIST_RETRIEVAL = [
"ArguAna",
"ClimateFEVER",
"CQADupstackRetrieval",
"DBPedia",
"FEVER",
"FiQA2018",
"HotpotQA",
"MSMARCO",
"NFCorpus",
"NQ",
"QuoraRetrieval",
"SCIDOCS",
"SciFact",
"Touche2020",
"TRECCOVID",
]
TASK_LIST_STS = [
"BIOSSES",
"SICK-R",
"STS12",
"STS13",
"STS14",
"STS15",
"STS16",
"STS17 (en-en)",
"STS22 (en)",
"STSBenchmark",
]
TASK_LIST_SUMMARIZATION = [
"SummEval",
]
TASK_LIST_EN = TASK_LIST_CLASSIFICATION + TASK_LIST_CLUSTERING + TASK_LIST_PAIR_CLASSIFICATION + TASK_LIST_RERANKING + TASK_LIST_RETRIEVAL + TASK_LIST_STS + TASK_LIST_SUMMARIZATION
TASK_TO_TASK_LIST = {}
def make_clickable_model(model_name):
# Remove user from model name
model_name_show = " ".join(model_name.split("/")[1:])
link = "https://huggingface.co/" + model_name
return (
f'<a target="_blank" style="text-decoration: underline" href="{link}">{model_name_show}</a>'
)
TASK_TO_METRIC = {
"BitextMining": "f1",
"Clustering": "v_measure",
"Classification": "accuracy",
"PairClassification": "cos_sim_ap",
"Reranking": "map",
"Retrieval": "ndcg_at_10",
"STS": "cos_sim_spearman",
"Summarization": "cos_sim_spearman",
}
def get_mteb_data(tasks=["Clustering"], metric="v_measure", langs=[], cast_to_str=True, task_to_metric=TASK_TO_METRIC):
api = HfApi()
models = api.list_models(filter="mteb")
df_list = []
for model in models:
readme_path = hf_hub_download(model.modelId, filename="README.md")
meta = metadata_load(readme_path)
# meta['model-index'][0]["results"] is list of elements like:
# {
# "task": {"type": "Classification"},
# "dataset": {
# "type": "mteb/amazon_massive_intent",
# "name": "MTEB MassiveIntentClassification (nb)",
# "config": "nb",
# "split": "test",
# },
# "metrics": [
# {"type": "accuracy", "value": 39.81506388702084},
# {"type": "f1", "value": 38.809586587791664},
# ],
# },
# Use "get" instead of dict indexing to skip incompat metadata instead of erroring out
#if langs is None:
task_results = [sub_res for sub_res in meta["model-index"][0]["results"] if (sub_res.get("task", {}).get("type", "") in tasks) and (sub_res.get("dataset", {}).get("config", "default") in ("default", *langs))]
out = [{res["dataset"]["name"].replace("MTEB ", ""): [round(score["value"], 2) for score in res["metrics"] if score["type"] == task_to_metric.get(res["task"]["type"])][0]} for res in task_results]
#else:
# Multilingual
# out = list(
# map(
# lambda x: {
# x["dataset"]["name"].replace("MTEB ", ""): round(
# list(filter(lambda x: x["type"] == metric, x["metrics"]))[0]["value"], 2
# )
# },
# filter(
# lambda x: (x.get("task", {}).get("type", "") in tasks)
# and (x.get("dataset", {}).get("config", "") in ("default", *langs)),
# meta["model-index"][0]["results"],
# ),
# )
# )
out = {k: v for d in out for k, v in d.items()}
out["Model"] = make_clickable_model(model.modelId)
df_list.append(out)
df = pd.DataFrame(df_list)
# Put 'Model' column first
cols = sorted(list(df.columns))
cols.insert(0, cols.pop(cols.index("Model")))
df = df[cols]
# df.insert(1, "Average", df.mean(axis=1, skipna=False))
df.fillna("", inplace=True)
if cast_to_str:
return df.astype(str) # Cast to str as Gradio does not accept floats
return df
DATA_OVERALL = get_mteb_data(
tasks=[
"Classification",
"Clustering",
"PairClassification",
"Reranking",
"Retrieval",
"STS",
"Summarization",
],
langs=["en", "en-en"],
cast_to_str=False
)
DATA_OVERALL.insert(1, "Average", DATA_OVERALL[TASK_LIST_EN].mean(axis=1, skipna=False))
DATA_OVERALL.insert(2, "Classification Average", DATA_OVERALL[TASK_LIST_CLASSIFICATION].mean(axis=1, skipna=False))
DATA_OVERALL.insert(3, "Clustering Average", DATA_OVERALL[TASK_LIST_CLUSTERING].mean(axis=1, skipna=False))
DATA_OVERALL.insert(4, "Pair Classification Average", DATA_OVERALL[TASK_LIST_PAIR_CLASSIFICATION].mean(axis=1, skipna=False))
DATA_OVERALL.insert(5, "Reranking Average", DATA_OVERALL[TASK_LIST_RERANKING].mean(axis=1, skipna=False))
DATA_OVERALL.insert(6, "Retrieval Average", DATA_OVERALL[TASK_LIST_RETRIEVAL].mean(axis=1, skipna=False))
DATA_OVERALL.insert(7, "STS Average", DATA_OVERALL[TASK_LIST_STS].mean(axis=1, skipna=False))
DATA_OVERALL.insert(8, "Summarization Average", DATA_OVERALL[TASK_LIST_SUMMARIZATION].mean(axis=1, skipna=False))
DATA_OVERALL = DATA_OVERALL.round(2).astype(str)
DATA_CLASSIFICATION_EN = DATA_OVERALL[["Model"] + TASK_LIST_CLASSIFICATION]
DATA_CLUSTERING = DATA_OVERALL[["Model"] + TASK_LIST_CLUSTERING]
DATA_PAIR_CLASSIFICATION = DATA_OVERALL[["Model"] + TASK_LIST_PAIR_CLASSIFICATION]
DATA_RERANKING = DATA_OVERALL[["Model"] + TASK_LIST_RERANKING]
DATA_RETRIEVAL = DATA_OVERALL[["Model"] + TASK_LIST_RETRIEVAL]
DATA_STS_EN = DATA_OVERALL[["Model"] + TASK_LIST_STS]
DATA_SUMMARIZATION = DATA_OVERALL[["Model"] + TASK_LIST_SUMMARIZATION]
DATA_OVERALL = DATA_OVERALL[["Model", "Average", "Classification Average", "Clustering Average", "Pair Classification Average", "Reranking Average", "Retrieval Average", "STS Average", "Summarization Average"]]
block = gr.Blocks()
with block:
gr.Markdown(
"""Leaderboard for XX most popular Blocks Event Spaces. To learn more and join, see <a href="https://huggingface.co/Gradio-Blocks" target="_blank" style="text-decoration: underline">Blocks Party Event</a>"""
)
with gr.Tabs():
with gr.TabItem("Overall"):
with gr.Row():
gr.Markdown("""Average Scores""")
with gr.Row():
data_overall = gr.components.Dataframe(
DATA_OVERALL,
datatype="markdown",
type="pandas",
col_count=(len(DATA_OVERALL.columns), "fixed"),
wrap=True,
)
with gr.TabItem("Classification"):
with gr.TabItem("English"):
with gr.Row():
gr.Markdown("""Leaderboard for Classification""")
with gr.Row():
data_classification_en = gr.components.Dataframe(
DATA_CLASSIFICATION_EN,
datatype="markdown",
type="pandas",
col_count=(len(DATA_CLASSIFICATION_EN.columns), "fixed"),
)
with gr.Row():
data_run = gr.Button("Refresh")
task_classification_en = gr.Variable(value="Classification")
metric_classification_en = gr.Variable(value="accuracy")
lang_classification_en = gr.Variable(value=["en"])
data_run.click(
get_mteb_data,
inputs=[
task_classification_en,
metric_classification_en,
lang_classification_en,
],
outputs=data_classification_en,
)
with gr.TabItem("Multilingual"):
with gr.Row():
gr.Markdown("""Multilingual Classification""")
with gr.Row():
data_classification = gr.components.Dataframe(
datatype=["markdown"] * 500,
type="pandas",
)
with gr.Row():
data_run = gr.Button("Refresh")
task_classification = gr.Variable(value="Classification")
metric_classification = gr.Variable(value="accuracy")
data_run.click(
get_mteb_data,
inputs=[task_classification, metric_classification],
outputs=data_classification,
)
with gr.TabItem("Clustering"):
with gr.Row():
gr.Markdown("""Leaderboard for Clustering""")
with gr.Row():
data_clustering = gr.components.Dataframe(
datatype=["markdown"] * 500,
type="pandas",
)
with gr.Row():
data_run = gr.Button("Refresh")
task_clustering = gr.Variable(value="Clustering")
metric_clustering = gr.Variable(value="v_measure")
data_run.click(
get_mteb_data,
inputs=[task_clustering, metric_clustering],
outputs=data_clustering,
)
with gr.TabItem("Retrieval"):
with gr.Row():
gr.Markdown("""Leaderboard for Retrieval""")
with gr.Row():
data_retrieval = gr.components.Dataframe(
datatype=["markdown"] * 500,
type="pandas",
)
with gr.Row():
data_run = gr.Button("Refresh")
task_retrieval = gr.Variable(value="Retrieval")
metric_retrieval = gr.Variable(value="ndcg_at_10")
data_run.click(
get_mteb_data, inputs=[task_retrieval, metric_retrieval], outputs=data_retrieval
)
with gr.TabItem("Reranking"):
with gr.Row():
gr.Markdown("""Leaderboard for Reranking""")
with gr.Row():
data_reranking = gr.components.Dataframe(
datatype=["markdown"] * 500,
type="pandas",
# col_count=(12, "fixed"),
)
with gr.Row():
data_run = gr.Button("Refresh")
task_reranking = gr.Variable(value="Reranking")
metric_reranking = gr.Variable(value="map")
data_run.click(
get_mteb_data, inputs=[task_reranking, metric_reranking], outputs=data_reranking
)
with gr.TabItem("STS"):
with gr.TabItem("English"):
with gr.Row():
gr.Markdown("""Leaderboard for STS""")
with gr.Row():
data_sts_en = gr.components.Dataframe(
datatype=["markdown"] * 500,
type="pandas",
)
with gr.Row():
data_run_en = gr.Button("Refresh")
task_sts_en = gr.Variable(value="STS")
metric_sts_en = gr.Variable(value="cos_sim_spearman")
lang_sts_en = gr.Variable(value=["en", "en-en"])
data_run.click(
get_mteb_data,
inputs=[task_sts_en, metric_sts_en, lang_sts_en],
outputs=data_sts_en,
)
with gr.TabItem("Multilingual"):
with gr.Row():
gr.Markdown("""Leaderboard for STS""")
with gr.Row():
data_sts = gr.components.Dataframe(
datatype=["markdown"] * 500,
type="pandas",
)
with gr.Row():
data_run = gr.Button("Refresh")
task_sts = gr.Variable(value="STS")
metric_sts = gr.Variable(value="cos_sim_spearman")
data_run.click(get_mteb_data, inputs=[task_sts, metric_sts], outputs=data_sts)
with gr.TabItem("Summarization"):
with gr.Row():
gr.Markdown("""Leaderboard for Summarization""")
with gr.Row():
data_summarization = gr.components.Dataframe(
datatype=["markdown"] * 500,
type="pandas",
)
with gr.Row():
data_run = gr.Button("Refresh")
task_summarization = gr.Variable(value="Summarization")
metric_summarization = gr.Variable(value="cos_sim_spearman")
data_run.click(
get_mteb_data,
inputs=[task_summarization, metric_summarization],
outputs=data_summarization,
)
# running the function on page load in addition to when the button is clicked
#block.load(
# get_mteb_data,
# inputs=[task_classification_en, metric_classification_en],
# outputs=data_classification_en,
# show_progress=False,
#)
block.load(
get_mteb_data,
inputs=[task_classification, metric_classification],
outputs=data_classification,
)
block.load(get_mteb_data, inputs=[task_clustering, metric_clustering], outputs=data_clustering)
block.load(get_mteb_data, inputs=[task_retrieval, metric_retrieval], outputs=data_retrieval)
block.load(get_mteb_data, inputs=[task_reranking, metric_reranking], outputs=data_reranking)
block.load(get_mteb_data, inputs=[task_sts, metric_sts], outputs=data_sts)
block.load(
get_mteb_data, inputs=[task_summarization, metric_summarization], outputs=data_summarization
)
block.launch()