Recently I have tried replicate.com for text classification
TIL that replicate.com deem every request as 1 second even if our request processing time only need milliseconds. On average 1 request only need around 30-45 ms.
Recently I have tried replicate.com for text classification
TIL that replicate.com deem every request as 1 second even if our request processing time only need milliseconds. On average 1 request only need around 30-45 ms.
This is the interface of their billing page before requests started
And after the requests were finished
Despite of that, replicate.com is a great platform for developers that need to integrate AI tools into their existing applications.
Test code:
https://gist.github.com/atrifat/e64b40475a7adf938f0563d301721049