Skip to content

[PaddleOCR_VL] 优化OCR提示逻辑 #4819

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

Open
megemini wants to merge 4 commits into
base: develop
Choose a base branch
from
Open

[PaddleOCR_VL] 优化OCR提示逻辑 #4819

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
b74a105
[PaddleOCR_VL] 优化OCR提示逻辑
megemini Dec 10, 2025
397ac92
[Fix]: 优化PaddleOCR_VL结果处理中的默认处理函数逻辑
megemini Dec 11, 2025
ed87c36
[Fix]: 简化PaddleOCR_VL结果处理中的默认处理函数逻辑
megemini Dec 11, 2025
b73e56f
[Fix]: 进一步简化PaddleOCR_VL结果处理中的默认处理函数格式
megemini Dec 11, 2025
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
21 changes: 14 additions & 7 deletions paddlex/inference/pipelines/paddleocr_vl/pipeline.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -253,12 +253,14 @@ def get_layout_parsing_results(
blocks.append(blocks_for_img)
for j, block in enumerate(blocks_for_img):
block_img = block["img"]
block_label = block["label"]
text_prompt = block["label"]
block_label = block["label"].lower()
if block_label not in image_labels and block_img is not None:
figure_token_map = {}
text_prompt = "OCR:"
drop_figures = []
if block_label == "table":
if block_label == "ocr":
text_prompt = "OCR:"
elif block_label == "table":
text_prompt = "Table Recognition:"
block_img, figure_token_map, drop_figures = (
tokenize_figure_of_table(
Expand Down Expand Up @@ -308,7 +310,7 @@ def get_layout_parsing_results(
for j, block in enumerate(blocks_for_img):
block_img = block["img"]
block_bbox = block["box"]
block_label = block["label"]
block_label = block["label"].lower()
block_content = ""
if curr_vlm_block_idx < len(vlm_block_ids) and vlm_block_ids[
curr_vlm_block_idx
Expand Down Expand Up @@ -447,12 +449,17 @@ def predict(
prompt_label = prompt_label if prompt_label else "ocr"
if prompt_label.lower() == "chart":
model_settings["use_chart_recognition"] = True
assert prompt_label.lower() in [
if prompt_label.lower() not in [
"ocr",
"formula",
"table",
"chart",
], f"Layout detection is disabled (use_layout_detection=False). 'prompt_label' must be one of ['ocr', 'formula', 'table', 'chart'], but got '{prompt_label}'."
]:
logging.warning(
f"Layout detection is disabled (use_layout_detection=False). "
f"'prompt_label' must be one of ['ocr', 'formula', 'table', 'chart'], "
f"but got '{prompt_label}'. Program will continue anyway."
)

def _process_cv(batch_data, new_batch_size=None):
if not new_batch_size:
Expand Down Expand Up @@ -510,7 +517,7 @@ def _process_cv(batch_data, new_batch_size=None):
"boxes": [
{
"cls_id": 0,
"label": prompt_label.lower(),
"label": prompt_label,
"score": 1,
"coordinate": [
0,
Expand Down
2 changes: 1 addition & 1 deletion paddlex/inference/pipelines/paddleocr_vl/result.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -510,7 +510,7 @@ def _to_markdown(self, pretty=True, show_formula_number=False) -> dict:
markdown_info["markdown_images"][block.image["path"]] = block.image[
"img"
]
handle_func = handle_funcs_dict.get(label, None)
handle_func = handle_funcs_dict.get(label, lambda block: block.content)
Copy link
Collaborator

@changdazhou changdazhou Dec 30, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这个位置的逻辑是对于页眉、页脚等方法要跳过展示，这里需要在前面加上判断model_setting是否关闭了版面，然后判断这个label是否在默认的原始dict中，如果不再则加入，建议merge最新develop分支代码后，在skip_markdown_label判断前加入

if (
show_formula_number
and (label == "display_formula" or label == "formula")
Expand Down