hbdat / iccv21_relational_direction Goto Github PK

Hello, thanks for your nice work. Is there any timeline to release the code?

Hello, hbdat:
Recently, I reproduce your work, but i encounter some problem, the chart is shown as below, the "fuxian" means my reproduced results and "paper" means the performance you reported in paper. The parameters I used are: epochs=10, batch_size=32, label_type='interaction', --partition train_1A, --mll_k_3 7, --mll_k_5 10, ----loc_k 10, and other parameters i don't change it. Looking forward to your reply, thanks!

Hi hdbat:
thanks for your nice work and I am very interested in this paper. However, after the code is finished trained, I get the following results.
In 1A setting.
"detail_human.csv & model_final.pt & partition_F1_3.csv & ranking_k_10.csv & stats.csv detail_hoi.csv & detail_object.csv & partition_AP.csv & partition_F1_5.csv & ranking_k_7.csv"

This is inconsistent with the results reported in the paper. Can you tell me how to reproduce it
Best

Hello! I was running the hico dataset and following the data preparation instruction. But the code cannot run and it reports missing the /data/hico_20160224_det/json/hoi_list.json file. I cannot find this file from the oringinal tar.gz file of hico dataset. Looking forward for your help, thx.