PaddleOCR 客户端有透出 input、output token数量吗

[liziwl]

PaddleOCR 客户端有透出 input、output token数量吗？

[aniruddhaadak80]

The title suggests a question about input and output token counting, but the body is too unclear to know which part of the pipeline you mean. It would help to clarify whether you are asking about OCR model internals, a downstream VLM or LLM step, or cost accounting for an end-to-end document pipeline.

If you share the exact model variant and where you expect token counts to appear, people can give a much more useful answer. Right now the missing context is not the implementation detail, it is which component you are actually trying to measure.