Unexpected "free-unknown" detections in PaddleNLP 2.3.0

See attached scan results. I scanned https://github.com/PaddlePaddle/PaddleNLP/archive/refs/tags/v2.3.0.tar.gz using a recent version of SCTK The results were generally quite impressive, but I was concerned about 274 "free-unknown" occurrences detected. These are generally of the pattern "# This file is distributed under the same license as the PaddleNLP package." Given that the Declared License of the project, apache-2.0, is definitely asserted properly and noted as such by the scan, it is misleading to refer to these as free-unknown. One might even argue (perhaps) that the developers are trying to do things properly with these lower level statements, which allow for flexibility in case the overall project license is changed at some point.

I imagine that generalized code to interpret the pattern "# This file is distributed under the same license as the PaddleNLP package." could be rather challenging, but it could be very valuable to make that attempt, especially since this kind of pattern could be the cause of other detections of the "unknown" type.

[PaddleNLP-2.3.0.tar.gz_scan.json.zip](https://github.com/nexB/scancode-toolkit/files/8718058/PaddleNLP-2.3.0.tar.gz_scan.json.zip)