Add batched inference possibility

[alonsosilvaallende]

By providing a clone attribute to the logits processors they get copied for all the batch:
I found it here in the vllm code (not documented I think):
https://github.com/vllm-project/vllm/blob/19dcc02a72e3ed52e3bf95aae44ea1f40ce42ea0/vllm/sampling_params.py#L537-L550

Solves #14

[aerdem4]

Great work!

Can you please

1. Sign your commits
2. Make sure flake8 doesn't complain
3. Increment package version so that I can update it on pypi after it is merged?

[alonsosilvaallende]

Thank you very much @aerdem4
I have fixed the issues. Let me know if something else is needed.

[aerdem4]

thanks @alonsosilvaallende all commits must have verified signatures for us to merge it. You can consider squashing them to one signed commit.

[alonsosilvaallende]

I tried to do what you mentioned but I don't think I succeded.

[aerdem4]

Now it has a merge conflict. You may get help from an LLM to make it work.

[alonsosilvaallende]

I tried with the help of claude. Does it work now?

[aerdem4]

Thank you. I just published the new version on pypi.