I am trying to run the NaQ pretrained model on just the validation set of the NLQ challenge.
When running the inference command bash ReLER/scripts/infer.sh 0 nlq_v2 test internvideo reler_internvideo_naq/video_tef-vlen600_internvideo/model_3448.t7, I am receiving the following error:
ValueError: data dir ./data/dataset/nlq_official_v2 does not exist
I am not really sure what should be going into that directory. Looking at the code, it seems to contain train/val/test.json files? Where do I get these files to be placed in this directory?
You can get nlq_official_v2 by following the pre-processing steps here. Note that nlq_train_v2.json and nlq_val_v2.json are the v2 release of the NLQ annotations.
I am trying to run the NaQ pretrained model on just the validation set of the NLQ challenge.
When running the inference command
bash ReLER/scripts/infer.sh 0 nlq_v2 test internvideo reler_internvideo_naq/video_tef-vlen600_internvideo/model_3448.t7
, I am receiving the following error:ValueError: data dir ./data/dataset/nlq_official_v2 does not exist
I am not really sure what should be going into that directory. Looking at the code, it seems to contain train/val/test.json files? Where do I get these files to be placed in this directory?