Lucas Daniel Velazquez M.
Lucas Daniel Velazquez M.
this just happened to me too
Looking over the code, I think this is caused by the generator thread not joining after the iteration stops, leaking its resources.
A solution might be adding a handle to join/terminate the thread when `GeneratorExit` and/or `StopIteration` is raised
Knowing the memory footprint for training would be great
For reference; I used an LLM to help me with the wording of this feature request.
I haven't tried with other APIs but at least openrouter seems to normalize all of the response formats to that of the OpenAI specification regardless of upstream model/provider (so long...
After reviewing other providers supported by ST, it appears they generally support some form of search, citations, or annotations. Each provider seems to have its own method for passing down...