Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add option to ignore casing in Includes eval template #655

Merged
merged 3 commits into from
Apr 12, 2023
Merged

Add option to ignore casing in Includes eval template #655

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
862c736
Add tests for ignore_case option for Includes eval template
trizko Apr 12, 2023
7cce6d0
Add option for ignoring casing in Includes eval template
trizko Apr 12, 2023
bdb9b93
Move ignore_case logic to utils.get_answer function
trizko Apr 12, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion evals/elsuite/basic/includes.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,12 +13,14 @@ def __init__(
self,
completion_fns: list[CompletionFn],
samples_jsonl: str,
ignore_case: bool = False,
*args,
**kwargs,
):
super().__init__(completion_fns, *args, **kwargs)
assert len(completion_fns) == 1, "Includes only supports one completion fn"
self.samples_jsonl = samples_jsonl
self.ignore_case = ignore_case

def eval_sample(self, sample: Any, *_):
prompt = sample["input"]
Expand All @@ -27,7 +29,9 @@ def eval_sample(self, sample: Any, *_):
)
sampled = result.get_completions()[0]

includes_answer = any([utils.get_answer(sampled, ref) for ref in sample["ideal"]])
includes_answer = any(
[utils.get_answer(sampled, ref, self.ignore_case) for ref in sample["ideal"]]
)
evals.record.record_metrics(accuracy=float(includes_answer))
return includes_answer

Expand Down
8 changes: 6 additions & 2 deletions evals/elsuite/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,8 +14,12 @@
)


def get_answer(text, answer_prompt):
idx = text.rfind(answer_prompt)
def get_answer(text, answer_prompt, ignore_case=False):
if ignore_case:
idx = text.lower().rfind(answer_prompt.lower())
else:
idx = text.rfind(answer_prompt)

if idx == -1:
return None
return text[idx + len(answer_prompt) :]
Expand Down
1 change: 1 addition & 0 deletions evals/registry/eval_sets/test-all.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@ test:
- test-match
- test-fuzzy-match
- test-includes
- test-includes-ignore-case
- coqa-match
- coqa-fact
- coqa-fact-expl
Expand Down
1 change: 1 addition & 0 deletions evals/registry/eval_sets/test-basic.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,3 +3,4 @@ test-basic:
- test-match
- test-fuzzy-match
- test-includes
- test-includes-ignore-case
10 changes: 10 additions & 0 deletions evals/registry/evals/test-basic.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,3 +25,13 @@ test-includes.s1.simple-v0:
class: evals.elsuite.basic.includes:Includes
args:
samples_jsonl: test_fuzzy_match/samples.jsonl

test-includes-ignore-case:
id: test-includes-ignore-case.s1.simple-v0
description: Example eval that uses fuzzy matching to score completions.
metrics: [accuracy]
test-includes-ignore-case.s1.simple-v0:
class: evals.elsuite.basic.includes:Includes
args:
samples_jsonl: test_fuzzy_match/samples.jsonl
ignore_case: true