kwrobel.eth
kwrobel.eth
In the main code, partial caching of openai models works for loglikelihood but not for generate. This PR should fix this.
And I confirm it works with `generate`.
Narrow view also does not work in Hero 11.
Please add `speed` variable available for formatting or at least add `.getDuration()` so the speed can be easily implemented as a custom payload. It is especially helpful when `total` is...
@saattrupdan Is it possible to add missing Finnish result?
@dosu I am using Agno framework with langfuse: 3.3.2 Model: claude-sonnet-4-20250514
I am sorry, I can't. I will try to use direct integration with Langfuse within Agno rather than using Opentelemetry way. Anyway (using Opentelemetry) it seems that Langfuse is doing...
I have used OpenLIT (https://langfuse.com/integrations/frameworks/agno-agents). Token counts looks better, however: 1. only input and output tokens are counted - there is no cache_writes and cache_reads 2. for LLM calls in...
Is it just a dump of all registered DNS names?