Text Classification
Laya
Safetensors
English
ledgar
lexglue
legal
contracts
calibrated
selective-prediction
Eval Results (legacy)
Instructions to use VibeHash/laya-ledgar with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Laya
How to use VibeHash/laya-ledgar with Laya:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
File size: 1,251 Bytes
779ff0b | 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 | """LEDGAR's option texts, settings and LexGLUE's two numbers on made-up predictions."""
from __future__ import annotations
import unittest
from layatasks.tasks import ledgar
from layatasks.tasks.ledgar import TASK, extra_scores
class OptionTests(unittest.TestCase):
def test_label_names_in_lower_case(self):
self.assertEqual(ledgar.option_text("Anti-Corruption Laws"), "anti-corruption laws")
self.assertEqual(TASK.question(["Assigns", "Assignments"])["criteria"], ["assigns", "assignments"])
def test_settings(self):
self.assertEqual((TASK.max_len, TASK.head_max_len, TASK.epochs, TASK.target_error), (768, 512, 4, 0.01))
self.assertEqual((TASK.loss, TASK.dev_rows), ("soft-ce", 2000))
class ScoreTests(unittest.TestCase):
def test_micro_and_macro_f1(self):
preds = [{"gold": g, "pred": q} for g, q in
[("notices", "notices"), ("notices", "notices"), ("waivers", "notices"), ("waivers", "waivers")]]
s = extra_scores(preds, None)
self.assertAlmostEqual(s["micro_f1"], 0.75)
# notices: tp 2, fp 1 -> 4/5. waivers: tp 1, fn 1 -> 2/3.
self.assertAlmostEqual(s["macro_f1"], (4 / 5 + 2 / 3) / 2)
if __name__ == "__main__":
unittest.main()
|