mindspore-lab / mindocr

A toolbox of ocr models and algorithms based on MindSpore
https://mindspore-lab.github.io/mindocr/
Apache License 2.0
210 stars 55 forks source link

FceNet评估报错:IndexError: index 1 is out of bounds for axis 0 with size 1 #544

Closed HustleOoo closed 8 months ago

HustleOoo commented 1 year ago

运行python tools/eval.py --config configs/det/fcenet/fce_icdar15.yaml时,发生报错如下:

Traceback (most recent call last): File "tools/eval.py", line 145, in main(config) File "tools/eval.py", line 135, in main measures = net_evaluator.eval() File "/home/mindocr-main/mindocr/utils/evaluator.py", line 153, in eval preds = self.postprocessor(preds, **data_info) File "/home/mindocr-main/mindocr/postprocess/det_base_postprocess.py", line 109, in call result = self.rescale(result, shape_list) File "/home/mindocr-main/mindocr/postprocess/det_base_postprocess.py", line 152, in rescale result[field][i] = self._rescale_polygons(sample, shape_list[i]) IndexError: index 1 is out of bounds for axis 0 with size 1

查看源码发现shape_list列表中只有一个元素,进行到第二个循环i=1时,发生越界,请问下这种情况如何解决?

panshaowu commented 8 months ago

@HustleOoo 您好,感谢您的反馈。该问题已经在PR #584 中修复。 该问题是shape_listresult['polys']的元素数量不一致引发的,已通过调整后处理结果中数据结构解决。 测试工程师已经在MindSpore r2.2.11上复验,可正常运行python tools/eval.py --config configs/det/fcenet/fce_icdar15.yaml脚本,请尝试拉取最新代码和运行。