Created by: ljx111

把deeplabv3的项目fork，换上自己的训练集，提示越界了，但是可视化中却没报错 还有想问一下TRAIN_CROP_SIZE，EVAL_CROP_SIZE和FIX_RESIZE_SIZE有什么关系吗？不太理解

#Device count: 1 W0411 11:40:04.963268 696 device_context.cc:259] Please NOTE: device: 0, CUDA Capability: 70, Driver API Version: 9.2, Runtime API Version: 9.0 W0411 11:40:04.967165 696 device_context.cc:267] device: 0, cuDNN Version: 7.3. load test model: work/model_ffc/final Traceback (most recent call last): File "./pdseg/eval.py", line 176, in main() File "./pdseg/eval.py", line 172, in main evaluate(cfg, **args.dict) File "./pdseg/eval.py", line 138, in evaluate conf_mat.calculate(pred, grts, masks) File "/home/aistudio/pdseg/metrics.py", line 47, in calculate shape=(self.num_classes, self.num_classes)) File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/scipy/sparse/compressed.py", line 57, in init other = self.class(coo_matrix(arg1, shape=shape)) File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/scipy/sparse/coo.py", line 198, in init self._check() File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/scipy/sparse/coo.py", line 285, in _check raise ValueError('row index exceeds matrix dimensions') ValueError: row index exceeds matrix dimensions