let's test the original model first:
Class Images Targets P R mAP F1: 100%|██████████| 32/32 [00:04<00:00, 7.45it/s]
all 128 929 0.0381 0.895 0.692 0.0722
......
Gamma value that less than 1.0166 are set to zero!
let's test the current model!
Class Images Targets P R mAP F1: 100%|██████████| 32/32 [00:00<00:00, 42.64it/s]
all 128 929 0 0 0 0
Number of channels has been reduced from 7584 to 3797
Prune ratio: 0.499
mAP of the 'pruned' model is 0.0000
let's test the original model first: Class Images Targets P R mAP F1: 100%|██████████| 32/32 [00:04<00:00, 7.45it/s] all 128 929 0.0381 0.895 0.692 0.0722 ......
Gamma value that less than 1.0166 are set to zero! let's test the current model! Class Images Targets P R mAP F1: 100%|██████████| 32/32 [00:00<00:00, 42.64it/s] all 128 929 0 0 0 0 Number of channels has been reduced from 7584 to 3797 Prune ratio: 0.499 mAP of the 'pruned' model is 0.0000
请问是数据集少,还是哪里出现的问题?