Batched inference API and support for float16 inference #279
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@salvaba94 I ran this batch inference demo but did not see any performance benefit to batching. Batch size 1 took 223ms on average, batch size 2 took 447ms on average, etc. It scaled linearly in the batch size. Did you observe the same behavior? Or were you able to get better throughput via batching? |
|
Hi @ZachOBrien, I've just checked it and yes, I see marginal improvements by using batching. Here are the results:
I guess the improvement depends on the GPU (this was tested with RTX 2060). |
|
did you encounter problems like #294 ? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This branch adds support for half precision inference and a batched inference API (BatchedModel). Additionally, it includes a short demo showing how to use this API.