Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
File size: 4,303 Bytes
ce5c604 6cdd0ad ce5c604 bba982c ce5c604 6cdd0ad ce5c604 a507ee8 ce5c604 6cdd0ad ce5c604 bcb8d03 adf26ec ce5c604 adf26ec ce5c604 5b935ed ce5c604 6cdd0ad ce5c604 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 |
import json
import os
from collections import defaultdict
import pandas as pd
from transformers import AutoConfig
from src.assets.hardcoded_evals import baseline
from src.display_models.get_model_metadata import apply_metadata
from src.display_models.read_results import get_eval_results_dicts, make_clickable_model
from src.display_models.utils import AutoEvalColumn, EvalQueueColumn, has_no_nan_values
def get_all_requested_models(requested_models_dir: str) -> set[str]:
depth = 1
file_names = []
users_to_submission_dates = defaultdict(list)
for root, _, files in os.walk(requested_models_dir):
current_depth = root.count(os.sep) - requested_models_dir.count(os.sep)
if current_depth == depth:
for file in files:
if not file.endswith(".json"): continue
with open(os.path.join(root, file), "r") as f:
info = json.load(f)
file_names.append(f"{info['model']}_{info['revision']}_{info['precision']}")
# Select organisation
if info["model"].count("/") == 0 or "submitted_time" not in info:
continue
organisation, _ = info["model"].split("/")
users_to_submission_dates[organisation].append(info["submitted_time"])
return set(file_names), users_to_submission_dates
def get_leaderboard_df(results_path: str, cols: list, benchmark_cols: list) -> pd.DataFrame:
all_data = get_eval_results_dicts(results_path)
# all_data.append(baseline)
apply_metadata(all_data) # Populate model type based on known hardcoded values in `metadata.py`
df = pd.DataFrame.from_records(all_data)
df = df.sort_values(by=[AutoEvalColumn.average.name], ascending=False)
df = df[cols].round(decimals=2)
# filter out if any of the benchmarks have not been produced
df = df[has_no_nan_values(df, benchmark_cols)]
return df
def get_evaluation_queue_df(save_path: str, cols: list) -> list[pd.DataFrame]:
entries = [entry for entry in os.listdir(save_path) if not entry.startswith(".")]
all_evals = []
for entry in entries:
if ".json" in entry:
file_path = os.path.join(save_path, entry)
with open(file_path) as fp:
data = json.load(fp)
data[EvalQueueColumn.model.name] = make_clickable_model(data["model"])
data[EvalQueueColumn.revision.name] = data.get("revision", "main")
all_evals.append(data)
elif ".md" not in entry:
# this is a folder
sub_entries = [e for e in os.listdir(f"{save_path}/{entry}") if not e.startswith(".")]
for sub_entry in sub_entries:
file_path = os.path.join(save_path, entry, sub_entry)
with open(file_path) as fp:
data = json.load(fp)
data[EvalQueueColumn.model.name] = make_clickable_model(data["model"])
data[EvalQueueColumn.revision.name] = data.get("revision", "main")
all_evals.append(data)
pending_list = [e for e in all_evals if e["status"] in ["PENDING", "RERUN"]]
running_list = [e for e in all_evals if e["status"] == "RUNNING"]
finished_list = [e for e in all_evals if e["status"].startswith("FINISHED") or e["status"] == "PENDING_NEW_EVAL"]
failed_list = [e for e in all_evals if e["status"] == "FAILED"]
df_pending = pd.DataFrame.from_records(pending_list, columns=cols)
df_running = pd.DataFrame.from_records(running_list, columns=cols)
df_finished = pd.DataFrame.from_records(finished_list, columns=cols)
df_failed = pd.DataFrame.from_records(failed_list, columns=cols)
return df_finished[cols], df_running[cols], df_pending[cols], df_failed[cols]
def is_model_on_hub(model_name: str, revision: str) -> bool:
try:
AutoConfig.from_pretrained(model_name, revision=revision, trust_remote_code=False)
return True, None
except ValueError:
return (
False,
"needs to be launched with `trust_remote_code=True`. For safety reason, we do not allow these models to be automatically submitted to the leaderboard.",
)
except Exception:
return False, "was not found on hub!"
|