MILVLG / bottom-up-attention.pytorch

A PyTorch reimplementation of bottom-up-attention models
Apache License 2.0
294 stars 76 forks source link

the code stopped when feature extraction #74

Closed KongYilun closed 2 years ago

KongYilun commented 3 years ago

HI ,the code is stoping at this: [09/24 13:23:39 detectron2]: Full config saved to ./output/config.yaml [09/24 13:23:39 d2.utils.env]: Using a generated random seed 39737149 Number of images: 26. 2021-09-24 13:23:41,216 INFO services.py:1265 -- View the Ray dashboard at http://127.0.0.1:8265 Number of GPUs: 1. 0%| | 0/26 00:00<?, ?it/s Number of images on split0: 26. (pid=2058) ResNet.make_stage(first_stride=) is deprecated! Use 'stride_per_block' or 'stride' instead. (pid=2058) ResNet.make_stage(first_stride=) is deprecated! Use 'stride_per_block' or 'stride' instead. (pid=2058) ResNet.make_stage(first_stride=) is deprecated! Use 'stride_per_block' or 'stride' instead.

gstoica27 commented 3 years ago

Have you tried removing all the ray calls? It worked for me after that (see #69).

rubencart commented 3 years ago

Same problem for me. @yuzcccc @Zoroaster97 @J1mL3e_ any solution?

Zoroaster97 commented 3 years ago

Hi, can you provide some Ray logs? You can check that on Ray dashboard (following the instruction: View the Ray dashboard at http://127.0.0.1:8265). And click the "View all logs".

image

leonodelee commented 2 years ago

Have you tried removing all the ray cal

Have you fixed this problem?