Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
gaoqiong
lm-evaluation-harness
Commits
c74e2761
Commit
c74e2761
authored
Dec 06, 2023
by
lintangsutawika
Browse files
reformat
parent
cc572624
Changes
155
Hide whitespace changes
Inline
Side-by-side
Showing
20 changed files
with
39 additions
and
39 deletions
+39
-39
lm_eval/api/metrics.py
lm_eval/api/metrics.py
+2
-2
lm_eval/api/task.py
lm_eval/api/task.py
+1
-1
lm_eval/tasks/arc/alternative_worlds/README.md
lm_eval/tasks/arc/alternative_worlds/README.md
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_01/a.yaml
lm_eval/tasks/arc/alternative_worlds/style_01/a.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_01/b.yaml
lm_eval/tasks/arc/alternative_worlds/style_01/b.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_01/c.yaml
lm_eval/tasks/arc/alternative_worlds/style_01/c.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_02/a.yaml
lm_eval/tasks/arc/alternative_worlds/style_02/a.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_02/b.yaml
lm_eval/tasks/arc/alternative_worlds/style_02/b.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_02/c.yaml
lm_eval/tasks/arc/alternative_worlds/style_02/c.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_03/a.yaml
lm_eval/tasks/arc/alternative_worlds/style_03/a.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_03/b.yaml
lm_eval/tasks/arc/alternative_worlds/style_03/b.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_03/c.yaml
lm_eval/tasks/arc/alternative_worlds/style_03/c.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_04/a.yaml
lm_eval/tasks/arc/alternative_worlds/style_04/a.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_04/b.yaml
lm_eval/tasks/arc/alternative_worlds/style_04/b.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_04/c.yaml
lm_eval/tasks/arc/alternative_worlds/style_04/c.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_05/a.yaml
lm_eval/tasks/arc/alternative_worlds/style_05/a.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_05/b.yaml
lm_eval/tasks/arc/alternative_worlds/style_05/b.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_05/c.yaml
lm_eval/tasks/arc/alternative_worlds/style_05/c.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_06/a.yaml
lm_eval/tasks/arc/alternative_worlds/style_06/a.yaml
+2
-2
lm_eval/tasks/arc/alternative_worlds/style_06/b.yaml
lm_eval/tasks/arc/alternative_worlds/style_06/b.yaml
+2
-2
No files found.
lm_eval/api/metrics.py
View file @
c74e2761
...
...
@@ -127,10 +127,10 @@ def brier_score(items): # This is a passthrough function
_p
=
np
.
array
(
p
)
_g
=
np
.
array
(
g
)
_g_one_hot
=
np
.
eye
(
len
(
_p
[
0
]))[
_g
]
average
+=
np
.
mean
(
np
.
sum
((
_p
-
_g_one_hot
)
**
2
,
axis
=
1
))
*
len
(
_g
)
average
+=
np
.
mean
(
np
.
sum
((
_p
-
_g_one_hot
)
**
2
,
axis
=
1
))
*
len
(
_g
)
total_size
+=
len
(
_g
)
return
average
/
total_size
return
average
/
total_size
@
register_metric
(
...
...
lm_eval/api/task.py
View file @
c74e2761
...
...
@@ -1020,7 +1020,7 @@ class ConfigurableTask(Task):
**
({
"perplexity"
:
ll
}
if
"perplexity"
in
use_metric
else
{}),
**
({
"acc"
:
int
(
is_greedy
)}
if
"acc"
in
use_metric
else
{}),
**
(
{
"brier_score"
:
(
0
,
[
prob_norm
])}
# Gold is Index 0
{
"brier_score"
:
(
0
,
[
prob_norm
])}
# Gold is Index 0
if
"brier_score"
in
use_metric
else
{}
),
...
...
lm_eval/tasks/arc/alternative_worlds/README.md
View file @
c74e2761
...
...
@@ -15,6 +15,6 @@ Answer types:
-
original option
-
just letter
-
letters + continuation
-
original option
-
original option
-
just letter
-
continuation
\ No newline at end of file
-
continuation
lm_eval/tasks/arc/alternative_worlds/style_01/a.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_01
group_alias
:
style_01
group_alias
:
style_01
task
:
arc_easy_01a
task_alias
:
a
doc_to_text
:
!function
../styles.template_01
doc_to_choice
:
!function
../styles.choice_01a
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_01a
lm_eval/tasks/arc/alternative_worlds/style_01/b.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_01
group_alias
:
style_01
group_alias
:
style_01
task
:
arc_easy_01b
task_alias
:
b
doc_to_text
:
!function
../styles.template_01
doc_to_choice
:
!function
../styles.choice_01b
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_01b
lm_eval/tasks/arc/alternative_worlds/style_01/c.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_01
group_alias
:
style_01
group_alias
:
style_01
task
:
arc_easy_01c
task_alias
:
c
doc_to_text
:
!function
../styles.template_01
doc_to_choice
:
!function
../styles.choice_01c
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_01c
lm_eval/tasks/arc/alternative_worlds/style_02/a.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_02
group_alias
:
style_02
group_alias
:
style_02
task
:
arc_easy_02a
task_alias
:
a
doc_to_text
:
!function
../styles.template_02
doc_to_choice
:
!function
../styles.choice_02a
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_02a
lm_eval/tasks/arc/alternative_worlds/style_02/b.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_02
group_alias
:
style_02
group_alias
:
style_02
task
:
arc_easy_02b
task_alias
:
b
doc_to_text
:
!function
../styles.template_02
doc_to_choice
:
!function
../styles.choice_02b
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_02b
lm_eval/tasks/arc/alternative_worlds/style_02/c.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_02
group_alias
:
style_02
group_alias
:
style_02
task
:
arc_easy_02c
task_alias
:
c
doc_to_text
:
!function
../styles.template_02
doc_to_choice
:
!function
../styles.choice_02c
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_02c
lm_eval/tasks/arc/alternative_worlds/style_03/a.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_03
group_alias
:
style_03
group_alias
:
style_03
task
:
arc_easy_03a
task_alias
:
a
doc_to_text
:
!function
../styles.template_03
doc_to_choice
:
!function
../styles.choice_03a
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_03a
lm_eval/tasks/arc/alternative_worlds/style_03/b.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_03
group_alias
:
style_03
group_alias
:
style_03
task
:
arc_easy_03b
task_alias
:
b
doc_to_text
:
!function
../styles.template_03
doc_to_choice
:
!function
../styles.choice_03b
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_03b
lm_eval/tasks/arc/alternative_worlds/style_03/c.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_03
group_alias
:
style_03
group_alias
:
style_03
task
:
arc_easy_03c
task_alias
:
c
doc_to_text
:
!function
../styles.template_03
doc_to_choice
:
!function
../styles.choice_03c
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_03c
lm_eval/tasks/arc/alternative_worlds/style_04/a.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_04
group_alias
:
style_04
group_alias
:
style_04
task
:
arc_easy_04a
task_alias
:
a
doc_to_text
:
!function
../styles.template_04
doc_to_choice
:
!function
../styles.choice_04a
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_04a
lm_eval/tasks/arc/alternative_worlds/style_04/b.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_04
group_alias
:
style_04
group_alias
:
style_04
task
:
arc_easy_04b
task_alias
:
b
doc_to_text
:
!function
../styles.template_04
doc_to_choice
:
!function
../styles.choice_04b
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_04b
lm_eval/tasks/arc/alternative_worlds/style_04/c.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_04
group_alias
:
style_04
group_alias
:
style_04
task
:
arc_easy_04c
task_alias
:
c
doc_to_text
:
!function
../styles.template_04
doc_to_choice
:
!function
../styles.choice_04c
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_04c
lm_eval/tasks/arc/alternative_worlds/style_05/a.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_05
group_alias
:
style_05
group_alias
:
style_05
task
:
arc_easy_05a
task_alias
:
a
doc_to_text
:
!function
../styles.template_05
doc_to_choice
:
!function
../styles.choice_05a
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_05a
lm_eval/tasks/arc/alternative_worlds/style_05/b.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_05
group_alias
:
style_05
group_alias
:
style_05
task
:
arc_easy_05b
task_alias
:
b
doc_to_text
:
!function
../styles.template_05
doc_to_choice
:
!function
../styles.choice_05b
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_05b
lm_eval/tasks/arc/alternative_worlds/style_05/c.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_05
group_alias
:
style_05
group_alias
:
style_05
task
:
arc_easy_05c
task_alias
:
c
doc_to_text
:
!function
../styles.template_05
doc_to_choice
:
!function
../styles.choice_05c
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_05c
lm_eval/tasks/arc/alternative_worlds/style_06/a.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_06
group_alias
:
style_06
group_alias
:
style_06
task
:
arc_easy_06a
task_alias
:
a
doc_to_text
:
!function
../styles.template_06
doc_to_choice
:
!function
../styles.choice_06a
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_06a
lm_eval/tasks/arc/alternative_worlds/style_06/b.yaml
View file @
c74e2761
include
:
../_arc_easy_alt_yaml
group
:
arc_easy_06
group_alias
:
style_06
group_alias
:
style_06
task
:
arc_easy_06b
task_alias
:
b
doc_to_text
:
!function
../styles.template_06
doc_to_choice
:
!function
../styles.choice_06b
\ No newline at end of file
doc_to_choice
:
!function
../styles.choice_06b
Prev
1
2
3
4
5
…
8
Next
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment