In gemini_translator.py, I see max_output_tokens is currently defined. Is it generally recommended to keep this set, or would the default behavior be preferable (e.g. gemini 3.1 flash lite have the default limit of 65536)?
I’m looking into max_output_tokens because I've noticed that full-chapter translations are occasionally being truncated when processed in a single API call.
In
gemini_translator.py, I see max_output_tokens is currently defined. Is it generally recommended to keep this set, or would the default behavior be preferable (e.g. gemini 3.1 flash lite have the default limit of 65536)?I’m looking into
max_output_tokensbecause I've noticed that full-chapter translations are occasionally being truncated when processed in a single API call.