Very high response time OCR Read API

Question

We are using OCR read API ( "vision/v3.2/read/analyze") to get the OCR results for a given image.

We were expecting a response time of less then 4 sec, however we are getting a response time of about 30 sec at an average.

we are using S1 pricing tier and we are sending 1 image at a time in a loop.

Code we are using is as below

with open(file_name, "rb") as image_stream:
data=image_stream.read()
response = requests.post(url,headers=headers,data=data)
analysis1 = {}

print("ocr-start",datetime.now())
while not "analyzeResult" in analysis1:
response_final = requests.get(response.headers["Operation-Location"], headers=headers)
analysis1 = response_final.json()
time.sleep(0.1)

Please help us, what are we doing wrong ..

Answer 1

@NeuranceAI Technologies Thanks for the question. We have a storage connection issue happened on 1/25 on CentralUS region.

or This is highly related to the complexity of document. Can you try using the preview model? Add this query parameter "model-version=2021-09-30-preview" to Read POST, and it will target to newer model, which has latency improvement.

------------------------------

• Please don't forget to click on or upvote button whenever the information provided helps you. Original posters help the community find answers faster by identifying the correct answer. Here is how
• Want a reminder to come back and check responses? Here is how to subscribe to a notification
• If you are interested in joining the VM program and help shape the future of Q&A: Here is how you can be part of Q&A Volunteer Moderators