Hi! Thanks for the excellent work.
I am trying to do inference with the checkpoint provided. I am using the command "main.py --model init --dataset htm-370k --batch_size 128 --use_text_pos_enc 0 --test [path]/TemporalAlignNet/ckpt/cotrain_word2vec_htm-370k_len64_e6d6.pth --inference 1".
Is this the correct way to do inference? Since I don't see where to pass in the (caption) sentences and video file. Also I am having the error "ModuleNotFoundError: No module named 'eval.inference_zeroshot_align'". In fact I didn't see the file inference_zeroshot_align.py in the eval folder at all.
How should this be solved? thanks!
Hi! Thanks for the excellent work.
I am trying to do inference with the checkpoint provided. I am using the command "main.py --model init --dataset htm-370k --batch_size 128 --use_text_pos_enc 0 --test [path]/TemporalAlignNet/ckpt/cotrain_word2vec_htm-370k_len64_e6d6.pth --inference 1".
Is this the correct way to do inference? Since I don't see where to pass in the (caption) sentences and video file. Also I am having the error "ModuleNotFoundError: No module named 'eval.inference_zeroshot_align'". In fact I didn't see the file inference_zeroshot_align.py in the eval folder at all.
How should this be solved? thanks!