File size: 2,000 Bytes
d9c0f48 f07d235 12efa10 ff92afe 12efa10 ff92afe f07d235 ff92afe 12efa10 f07d235 ff92afe 12efa10 1afc4a7 ff92afe 12efa10 ff92afe 12efa10 ff92afe | 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 | from dataclasses import dataclass, make_dataclass, field
from src.about import EvalDimensions
def fields(raw_class):
# This helper looks at the class __dict__, so we must ensure
# the objects are actually sitting on the class.
return [v for k, v in raw_class.__dict__.items() if k[:2] != "__" and k[-2:] != "__"]
@dataclass
class ColumnContent:
name: str
type: str
displayed_by_default: bool
hidden: bool = False
never_hidden: bool = False
column_data = {
"rank": ColumnContent("Rank", "str", True, False),
"model_source": ColumnContent("Source", "str", True, False),
"model_category": ColumnContent("Size", "str", True, False),
"model": ColumnContent("Model Name", "markdown", True, never_hidden=True),
"average_score": ColumnContent("Benchmark Score (0-10)", "number", True),
}
for eval_dim in EvalDimensions:
is_visible = eval_dim.value.metric in ["speed", "contamination_score"]
column_data[eval_dim.name] = ColumnContent(eval_dim.value.col_name, "number", is_visible)
auto_eval_column_dict = []
for name, content in column_data.items():
# We use a closure to capture the specific 'content' for each field
auto_eval_column_dict.append((
name,
ColumnContent,
field(default_factory=lambda c=content: c)
))
# Create the dynamic dataclass
AutoEvalColumn = make_dataclass("AutoEvalColumn", auto_eval_column_dict, frozen=True)
for name, content in column_data.items():
setattr(AutoEvalColumn, name, content)
## For the queue columns
@dataclass(frozen=True)
class EvalQueueColumn:
model = ColumnContent("model", "markdown", True)
revision = ColumnContent("revision", "str", True)
status = ColumnContent("status", "str", True)
# Column selection
COLS = [c.name for c in fields(AutoEvalColumn) if not c.hidden]
EVAL_COLS = [c.name for c in fields(EvalQueueColumn)]
EVAL_TYPES = [c.type for c in fields(EvalQueueColumn)]
BENCHMARK_COLS = [t.value.col_name for t in EvalDimensions] |