I see that llama-stack is becoming very powerful set of tool which sits on top of LLM model.
inference, memory, agent, scoring, eval etc can be used via APIs
but for finetuning, either today we have to go out of llama stack ecosystem to torchtune or other libraries
i strongly believe if we add API for finetuning (atleast to start with single node meta-reference), llama-stack will become one stop s
ashwinb
commented
2 weeks ago
🚀 The feature, motivation and pitch
I see that llama-stack is becoming very powerful set of tool which sits on top of LLM model. inference, memory, agent, scoring, eval etc can be used via APIs
but for finetuning, either today we have to go out of llama stack ecosystem to torchtune or other libraries
i strongly believe if we add API for finetuning (atleast to start with single node meta-reference), llama-stack will become one stop s
Alternatives
No response
Additional context
No response