search

MCG-NJU / AdaMixer

[CVPR 2022 Oral] AdaMixer: A Fast-Converging Query-Based Object Detector
MIT License
236 stars 24 forks source link

RuntimeError: output with shape [16, 300, 300] doesn't match the broadcast shape [1, 16, 300, 300] #12

Closed sanmulab closed 2 years ago

sanmulab commented 2 years ago

I don't know if it is a problem with my environment, I just changed the dataset path to my own, but the error is reported. My torch==1.3.1. error: File "/home/ubuntu/.conda/envs/mmlab/lib/python3.7/site-packages/torch/nn/functional.py", line 3352, in multi_head_attention_forward attn_output_weights += attn_mask RuntimeError: output with shape [16, 300, 300] doesn't match the broadcast shape [1, 16, 300, 300]

sebgao commented 2 years ago

What is your MMCV version? You can view it by pip show mmcv or pip show mmcv-full.

sebgao commented 2 years ago

By the way, we recommend to install pytorch>=1.5

sanmulab commented 2 years ago

Thanks! Problem solved, I used the new environment for training.

sanmulab commented 2 years ago

Test problem: When the model starts to evaluate, an error is reported: AttributeError: 'InitialQueryGenerator' object has no attribute 'aug_test_rpn'. Why is this? The default test configuration I use: test_cfg=dict(rpn=None, rcnn=dict(max_per_img=300)

sebgao commented 2 years ago

Sorry, multi-scale testing is not currently supported by AdaMixer. Please use the single scale testing data loading protocol.