Classic298

Results 219 comments of Classic298

@krrishdholakia what i also noticed: ![Image](https://github.com/user-attachments/assets/f5e848e4-92c0-489d-8c4f-7ad1a29562bf) This is the response usage metadata object for a request to gemini 2.5 pro. Calculate the costs yourself. You will notice, LiteLLM did not...

yeah same. I also never get cache details which is why i asked here if LiteLLM supports it yet

> For prompts under 5K tokens, implicit caching was very inconsistent—even small changes in user input (like tweaking a character or a variable) could randomly cause cache hits or misses....

@ishaan-jaff litellm[proxy] python package I'll try to get you the stacktrace tomorrow.

@krrishdholakia @ishaan-jaff No I do not build from docker, I installed litellm[proxy] via pip ``` May 13 13:42:58 systemd[1]: Started litellm.service - LiteLLM AI Gateway. May 13 13:43:01 litellm[359962]: Traceback...

@tjbck any chance to at least get the fix in by the next version for duplicate content checks?

RE: https://github.com/open-webui/open-webui/issues/17088 and https://github.com/open-webui/open-webui/discussions/18119

@tjbck TLDR: - Fixes very old existing issue of supposed duplicate content where there is none - Adds feature to attach website to knowledge bases - PR is tested multiple...