NousResearch/atropos
1
0
Fork 0
mirror of https://github.com/NousResearch/atropos.git synced 2026-04-19 12:57:58 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

[pre-commit.ci] auto fixes from pre-commit.com hooks

Browse source 
for more information, see https://pre-commit.ci
This commit is contained in:
pre-commit-ci[bot] 2026-01-12 23:29:15 +00:00
parent 22884d2bf7
commit d84e3c70b7
16 changed files with 270 additions and 143 deletions
Download patch file Download diff file Expand all files Collapse all files

40
environments/eval_environments/ocrbench_environment.py
View file

@ -17,16 +17,16 @@ class OCRBench(EvalBase):
# Categories and their scoring
CATEGORIES = [
'Regular Text Recognition',
'Irregular Text Recognition',
'Artistic Text Recognition',
'Handwriting Recognition',
'Digit String Recognition',
'Non-Semantic Text Recognition',
'Scene Text-centric VQA',
'Doc-oriented VQA',
'Key Information Extraction',
'Handwritten Mathematical Expression Recognition',
"Regular Text Recognition",
"Irregular Text Recognition",
"Artistic Text Recognition",
"Handwriting Recognition",
"Digit String Recognition",
"Non-Semantic Text Recognition",
"Scene Text-centric VQA",
"Doc-oriented VQA",
"Key Information Extraction",
"Handwritten Mathematical Expression Recognition",
]
def setup_data(self) -> list:
@ -65,10 +65,12 @@ class OCRBench(EvalBase):
content = []
if image_base64:
content.append({
"type": "image_url",
"image_url": {"url": f"data:image/png;base64,{image_base64}"},
})
content.append(
{
"type": "image_url",
"image_url": {"url": f"data:image/png;base64,{image_base64}"},
}
)
content.append({"type": "text", "text": prompt})
return [{"role": "user", "content": content}]
@ -77,16 +79,16 @@ class OCRBench(EvalBase):
"""Category-specific scoring for OCR tasks."""
predict = prediction.strip()
if category == 'Handwritten Mathematical Expression Recognition':
predict_clean = predict.replace('\n', ' ').replace(' ', '')
if category == "Handwritten Mathematical Expression Recognition":
predict_clean = predict.replace("\n", " ").replace(" ", "")
for answer in answers:
answer_clean = answer.strip().replace('\n', ' ').replace(' ', '')
answer_clean = answer.strip().replace("\n", " ").replace(" ", "")
if answer_clean in predict_clean:
return True
else:
predict_lower = predict.lower().replace('\n', ' ')
predict_lower = predict.lower().replace("\n", " ")
for answer in answers:
answer_lower = answer.lower().strip().replace('\n', ' ')
answer_lower = answer.lower().strip().replace("\n", " ")
if answer_lower in predict_lower:
return True