Legal NLP tasks on Swiss data #1032
JoelNiklaus
merged
16 commits into
huggingface:main
from
rolshoven:community_task_slds
Legal NLP tasks on Swiss data
83c8a079
refactor: split Swiss legal multilingual tasks into modular package
8d19996d
refactor: update higher_is_better type in MetricGrouping
4032ce45
refactor: Updated prompts and implementation to match the latest SwiL…
ca7e8342
refactor: Enhance COMET and GEMBA metric loading with error handling
606417b8
Add Gemba dependency for Swiss legal evaluations and remove `suite` p…
8c449c2c
Fix batched metric aggregation for grouped metric names
9113d537
Fixed missing system prompt
6d3fdf3b
Judge models now are used through OpenRouter
bba0e914
fix reasoning model token handling when max_tokens is unset
cc22ecb6
chore: trigger PR update
a4e0ba13
Merge branch 'main' into community_task_slds
406953fd
NathanHB
approved these changes
on 2026-05-20
fix: return raw score for BLEU, CHRF, and TER metrics instead of scal…
052586aa
fix: replaced accidental default value assignment with intended type …
fa341ac9
fix: add error handling for unsupported languages in Swiss Landmark D…
011c4047
fix: avoid huge negative BERTScore from baseline rescaling
df00dc16
Login to write a write a comment.
Login via GitHub