inference

[LilyTheBear]

Hi,
Do you have a sample inference code to load the model, pre-process video and text, and get the similarity score ?

Thanks !

[LeeYN-43]

Hi,
You can run bash tools/dist_test.sh [config file path] [model path] [GPU number] --eval recall_for_video_text_retrieval
This linux command will load the model and run it on the test dataset specified in your config file. You can get video-text similarity score in the "recall_for_video_text_retrieval" function at mmaction/core/evaluation/accuracy.py

Hope this can help you : )