Max Goltzsche
Max Goltzsche
That looks awesome! I'll try to prepare a khelm example with it during the weekend.
That is true! For this to work with e.g. khelm requires to be able to declare also non-yaml files input. I had another issue in mind that could solve this...
Is there any progress on this end by any chance? Currently the lack of prompt caching prevents having a longer conversation history in real-time apps and results in higher latency...
I found a corresponding issue within the llama.cpp project: https://github.com/ggml-org/llama.cpp/issues/17118
A concrete use case for this is [khelm](https://github.com/mgoltzsche/khelm). It requires to provide a local Helm chart (including Helm templates that have the .yaml extension but don't contain valid yaml) as...
> Most use cases we've seen can be formulated as running a priviledged function once instead of repeatedly as part of the pipeline. That's true and still allows declarative usage...
I remember also having seen that error not long ago somewhere but unfortunately I missed to capture it within a bug ticket and now I cannot reproduce it locally on...