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
40027bca
Commit
40027bca
authored
Jan 02, 2025
by
Baber
Browse files
nit
parent
2de3ffdf
Changes
2
Show whitespace changes
Inline
Side-by-side
Showing
2 changed files
with
4 additions
and
1 deletion
+4
-1
lm_eval/tasks/llama3/instruct/mgsm/mgsm_chat.yaml
lm_eval/tasks/llama3/instruct/mgsm/mgsm_chat.yaml
+3
-1
lm_eval/tasks/llama3/instruct/mgsm/mgsm_chat_template
lm_eval/tasks/llama3/instruct/mgsm/mgsm_chat_template
+1
-0
No files found.
lm_eval/tasks/llama3/instruct/mgsm/mgsm_chat.yaml
View file @
40027bca
group
:
mgsm_chat
group
:
mgsm_chat
group_alias
:
m
mlu
(llama)
group_alias
:
m
gsm
(llama)
task
:
task
:
-
mgsm_chat_bn
-
mgsm_chat_bn
-
mgsm_chat_de
-
mgsm_chat_de
...
@@ -14,6 +14,8 @@ task:
...
@@ -14,6 +14,8 @@ task:
-
mgsm_chat_zh
-
mgsm_chat_zh
aggregate_metric_list
:
aggregate_metric_list
:
-
metric
:
exact_match
-
metric
:
exact_match
aggregation
:
mean
weight_by_size
:
True
weight_by_size
:
True
filter_list
:
[
flexible-extract
,
strict-match
]
metadata
:
metadata
:
version
:
0
version
:
0
lm_eval/tasks/llama3/instruct/mgsm/mgsm_chat_template
View file @
40027bca
...
@@ -26,6 +26,7 @@ filter_list:
...
@@ -26,6 +26,7 @@ filter_list:
- name: "strict-match"
- name: "strict-match"
filter:
filter:
- function: "regex"
- function: "regex"
group_select: -1
regex_pattern: "(?:Answer|Réponse|Antwort|Ответ|Respuesta|答え|Jibu|答案|คำตอบ|సమాధానం|উত্তর): (\\-?[0-9\\.\\,]+)"
regex_pattern: "(?:Answer|Réponse|Antwort|Ответ|Respuesta|答え|Jibu|答案|คำตอบ|సమాధానం|উত্তর): (\\-?[0-9\\.\\,]+)"
- function: remove_whitespace
- function: remove_whitespace
- function: take_first
- function: take_first
...
...
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