Closed
Description
🚀 The feature, motivation and pitch
In my use case I'm running many offline evaluations of models, so the terminal quickly gets cluttered with vLLM progress bars. This is due to the progress bar in the LLM._run_engine
method as well as the newer "Adding requests" progress bar in the LLM._add_requests_with_fixed_progress_bars
method.
I propose adding a new Boolean argument, tqdm_leave_pbar
, defaulting to True
, consistent with the current behaviour, to all functions with the use_tqdm
argument.
Alternatives
I'm currently dealing with this myself in a hacky way by replacing the above-mentioned methods of LLM
with copied ones where I have set leave=False
manually. This is of course not sustainable, as I don't get any updates to these methods.
Additional context
No response
Before submitting a new issue...
- Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.