lighteval
Add metrics as functions
#214
Merged

Add metrics as functions #214

clefourrier merged 27 commits into main from metrics_as_fn
clefourrier
hynky1999 add function prompt assigment
58fee656
hynky1999 add json casting
57d40f96
hynky1999 fix ruff setting + fmt
cb8be215
clefourrier Merge branch 'main' into function_prompts
f275f600
clefourrier Merge branch 'main' into function_prompts
cde6c04f
clefourrier replaced json tasks by python tasks, step 1
ced2945e
clefourrier wip
82e98152
clefourrier Merge branch 'main' into simplify_task_system
a6aa1335
clefourrier simplification part 1
c5f428cc
clefourrier fix extended tasks + typo
4410c237
clefourrier fix
c93a2fa0
clefourrier fix nanotron example
96767564
clefourrier small fix
b84f006c
clefourrier Merge branch 'main' into function_prompts
c656d64a
clefourrier Merge branch 'simplify_task_system' into hynek_function
770f67e8
clefourrier now use function, not string, to pass prompts in examples
d43ffacb
clefourrier moved everyone to function calling
e10a84ce
clefourrier LightevalTask now only takes functions
c927d143
clefourrier removed templated type which messed up the test suite
e4182b43
clefourrier last fix + doc udpate
9f518ade
clefourrier moving every metric launcher to a metric
b3cfbc2a
clefourrier typo fix + fixes json encoder to fit metric storage
417a7aab
clefourrier clefourrier changed the base branch from main to hynek_function 1 year ago
clefourrier fix typo
129ba24f
clefourrier fix tests
7b614319
clefourrier clefourrier requested a review from NathanHB NathanHB 1 year ago
clefourrier clefourrier changed the base branch from hynek_function to main 1 year ago
clefourrier Merge branch 'main' into metrics_as_fn
f957a9bd
clefourrier Merge branch 'main' into metrics_as_fn
eedaa592
NathanHB
NathanHB commented on 2024-07-12
NathanHB Merge branch 'main' into metrics_as_fn
895c072f
NathanHB
NathanHB approved these changes on 2024-07-16
clefourrier clefourrier merged aaf7e8af into main 1 year ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone