Ideal parameters when serving only one client?

[andysalerno]

Hi,

I'm using TGI to serve up an assistant, for which I'm the only user.

TGI might be a bit overkill for this, since it is geared toward efficiently serving many clients. But it's so easy to set up, and so stable, that it's working beautifully for this scenario :)

So my question is, if I know I am only serving one request at a time, what would be the ideal parameters?

I'm guessing these:

---validation-workers 1

--max-concurrent-requests 1

but I'm not sure if there are any other considerations or any other things I could do.