MSVD checkpoint

[avinashsai]

Hi,

Congrats on your amazing work! Can you please upload the MSVD checkpoint and steps for inference?

[jpthu17]

Thank you for your attention to our work.
The code and inference steps for MSVD can be found in our work EMCL (https://github.com/jpthu17/EMCL/tree/main/video_retrieval/EMCL-Net#train-on-msvd).
I have other things to do recently, so I will probably update this codebase in June and upload the MSVD checkpoint.

[avinashsai]

Thanks for your comment. When I run using the same MSVD dataset in ECML I get the following error:

[2023-05-06 23:11:45,935 tvr 540 INFO]: [finish] extract text+video feature [2023-05-06 23:11:45,964 tvr 573 INFO]: 2145 52 2145 52 [2023-05-06 23:11:45,964 tvr 577 INFO]: [start] calculate the similarity [2023-05-06 23:11:45,964 tvr 383 INFO]: [finish] map to main gpu [2023-05-06 23:11:45,975 tvr 397 INFO]: [finish] map to main gpu 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████| 34/34 [00:26<00:00, 1.28it/s] [2023-05-06 23:15:37,725 tvr 580 INFO]: [end] calculate the similarity [2023-05-06 23:15:37,725 tvr 583 INFO]: [start] compute_metrics [2023-05-06 23:15:37,725 tvr 585 INFO]: before reshape, sim matrix size: 2145 x 52 Traceback (most recent call last): File "main_retrieval.py", line 700, in <module> main() File "main_retrieval.py", line 663, in main scheduler, global_step, max_steps, val_dataloader) File "main_retrieval.py", line 341, in train_epoch R1 = eval_epoch(args, model, val_dataloader, args.device) File "main_retrieval.py", line 593, in eval_epoch t2v_sim_matrix_new = np.stack(tuple(t2v_sim_matrix_new), axis=0) File "<__array_function__ internals>", line 6, in stack File "/playpen/avinashm/anaconda3/envs/frozen/lib/python3.7/site-packages/numpy/core/shape_base.py", line 427, in stack raise ValueError('all input arrays must have the same shape') ValueError: all input arrays must have the same shape

[beatriceadel]

Hi, I am facing the same issue when trying to train on the MSVD dataset. I got the same errors as the message above.

[jpthu17]

Hi, I am facing the same issue when trying to train on the MSVD dataset. I got the same errors as the message above.

I'm sorry for not replying in time. I've been busy with other DDLs. I will fix the bug as soon as possible.

[adrianofragomeni]

I have the same error when training with MSVD. Could you pleasetry to solve the bug?