Share via

Hot to decrease response time for ReadAPI?

Mattia Durli 1 Reputation point
2021-02-05T16:02:02.953+00:00

We're using ReadAPI to read some simple labels with little information, the final JSON is about 5kb.
The uploaded JPG 100kb.
The response time varies between 0,5s (rarely) and 2s, more often around 1s.
We check result every 250ms.

Result quality is very good, but is a litlle bit slow.

Is there a way to improve the response time?
A more expensive tier with more computational resource?

We were thinking of implementing the docker version locally, but what time of computer/processor/memory would be necessary to compare with the cloud one?
Then we would have to go for a more powerful one.
Trying to understand if it is worth it to go for the docker local solution.

Foundry Tools
Foundry Tools

Formerly known as Azure AI Services or Azure Cognitive Services is a unified collection of prebuilt AI capabilities within the Microsoft Foundry platform

Azure Vision in Foundry Tools

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. GiftA-MSFT 11,181 Reputation points Moderator
    2021-02-09T02:12:15.843+00:00

    Hi, thanks for following up. There's no way to optimize the response time based on tier. You can experiment using containers to reduce latency.

    Was this answer helpful?

    0 comments

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.