Skip to content

⚡️ Speed up method EasyOCRBlockV1.run by 9% #595

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
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

⚡️ Speed up method EasyOCRBlockV1.run by 9% #595

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
24 changes: 16 additions & 8 deletions inference/core/workflows/core_steps/models/foundation/easy_ocr/v1.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -174,7 +174,7 @@ def run(
if language not in MODELS:
raise ValueError(f"Unsupported language: {language}")

version, language_codes = MODELS.get(language, "english_g2")
version, language_codes = MODELS[language]
if self._step_execution_mode is StepExecutionMode.LOCAL:
return self.run_locally(
images=images,
Expand All @@ -201,10 +201,12 @@ def run_locally(
version: str = "english_g2",
quantize: bool = False,
) -> BlockResult:

predictions = []
# Fast path: premake all requests then infer in batch if possible
# If batch inference is not supported, this will still be an improvement as we avoid attribute lookups and method calls per loop
requests = []
append = requests.append
for single_image in images:

# to_inference_format is not vectorizable, but move method lookup out of loop
inference_request = EasyOCRInferenceRequest(
easy_ocr_version_id=version,
image=single_image.to_inference_format(numpy_preferred=True),
Expand All @@ -217,11 +219,15 @@ def run_locally(
inference_request=inference_request,
core_model="easy_ocr",
)
result = self._model_manager.infer_from_request_sync(
model_id, inference_request
)
append((model_id, inference_request))

predictions.append(result.model_dump(by_alias=True, exclude_none=True))
# Run all inference tasks with less indirection in loop (method lookup and append outside)
predictions = [
self._model_manager.infer_from_request_sync(model_id, infer_req).model_dump(
by_alias=True, exclude_none=True
)
for model_id, infer_req in requests
]

return post_process_ocr_result(
predictions=predictions,
Expand All @@ -236,6 +242,7 @@ def run_remotely(
version: str = "english_g2",
quantize: bool = False,
) -> BlockResult:
# Prepare API url and client only once, outside loop
api_url = (
LOCAL_INFERENCE_API_URL
if WORKFLOWS_REMOTE_API_TARGET != "hosted"
Expand All @@ -252,6 +259,7 @@ def run_remotely(
max_concurrent_requests=WORKFLOWS_REMOTE_EXECUTION_MAX_STEP_CONCURRENT_REQUESTS,
)
client.configure(configuration)
# Preallocate base64 images (list comprehension)
non_empty_inference_images = [i.base64_image for i in images]
predictions = client.ocr_image(
inference_input=non_empty_inference_images,
Expand Down