Description & Motivation
Since the whole idea is prompting as a new programming language, meaning meaningful manipulation of the prompts.
I think your project is perfectly positioned to incorporate caching optimization: many AI providers such as OpenAI & Anthropic adopt some type of caching both boils down to placing repetative texts such as context at the beginning of the prompt.
While some vendors require explicit caching to be enabled such as Anthropic making it harder to use both AdalFlow & Caching.
Pitch
Since AdalFlow is a prompt manipulation platform, I think you can also take into account caching to not only improve quality of the reply but also reduce costs.
Alternatives
Manually do that.
Additional context
No response
Description & Motivation
Since the whole idea is prompting as a new programming language, meaning meaningful manipulation of the prompts.
I think your project is perfectly positioned to incorporate caching optimization: many AI providers such as OpenAI & Anthropic adopt some type of caching both boils down to placing repetative texts such as context at the beginning of the prompt.
While some vendors require explicit caching to be enabled such as Anthropic making it harder to use both AdalFlow & Caching.
Pitch
Since AdalFlow is a prompt manipulation platform, I think you can also take into account caching to not only improve quality of the reply but also reduce costs.
Alternatives
Manually do that.
Additional context
No response