search

JacobYuan7 / RLIPv2

[ICCV 2023] RLIPv2: Fast Scaling of Relational Language-Image Pre-training
Apache License 2.0
112 stars 3 forks source link

No module named 'MultiScaleDeformableAttention'. #20

Open Srisude opened 3 weeks ago

Srisude commented 3 weeks ago

Hi, when i run the folowing code in colab :!python3 /content/v-coco/RLIPv2/main.py --hoi --dataset_file vcoco --hoi_path /content/v-coco/data/annotations --coco_path /content/v-coco --output_dir /content/outputs --epochs 50 --batch_size 2 --num_queries I get the error that: import MultiScaleDeformableAttention as MSDA ModuleNotFoundError: No module named 'MultiScaleDeformableAttention'. ı found the module from https://github.com/fundamentalvision/Deformable-DETR but it doesnt work,where is the mistake, thank youu.

Srisude commented 3 weeks ago

Also I didnt work in a conda environment because conda was running slowly in colab( ı got cuda error on mac)and ı downloaded requirements one by one with pip, Is the above prob also related to this?

JacobYuan7 commented 3 weeks ago

@Srisude It seems that you haven't installed the package correctly. Can you check your environment to see if you have done so?