Show HN: SynapseKit – Async-native Python framework for LLM pipelines and agents(github.com)
github.com
Show HN: SynapseKit – Async-native Python framework for LLM pipelines and agents
https://github.com/SynapseKit/SynapseKit
4 コメント
https://www.reddit.com/r/LangChain/comments/1sm3g2q/i_tested...
Async is definitely the right approach for this, but how are you handling the event loop blocking if a local model takes way too long to generate a heavy sequence?