CVMI-Lab / PAConv

(CVPR 2021) PAConv: Position Adaptive Convolution with Dynamic Kernel Assembling on Point Clouds
Apache License 2.0
287 stars 40 forks source link

WARNING:root:NaN or Inf found in input tensor.? #29

Closed swzaaaaaaa closed 2 years ago

swzaaaaaaa commented 2 years ago

[2021-12-29 09:19:22,793 INFO train.py line 249 2152192] Epoch: [1/100][1900/2976] Data 0.001 (0.004) Batch 0.935 (0.982) Remain 80:39:57 Loss 1.0328 Main Loss 1.0328 Accuracy 0.5974. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. WARNING:root:NaN or Inf found in input tensor. [2021-12-29 09:20:59,144 INFO train.py line 249 2152192] Epoch: [1/100][2000/2976] Data 0.001 (0.004) Batch 0.911 (0.981) Remain 80:33:45 Loss nan Main Loss nan Accuracy 0.2679. INFO:main-logger:Epoch: [1/100][2000/2976] Data 0.001 (0.004) Batch 0.911 (0.981) Remain 80:33:45 Loss nan Main Loss nan Accuracy 0.2679.

swzaaaaaaa commented 2 years ago

请问,你有遇到过这个问题吗?