question about the steps4 on holicity_dataset

[DennisJcy]

Hi,i follow your guide and face a mistake in steps 4,and im sure that i run the step2 and 3 to generate the gt data according your advise in #19 . Can you give me some advice?thank you very much!

[Info] Training mode Output path: ./experiments/sold2_hair Initializing dataset... [Info] Initializing Holicity dataset... Found filename cache holicity_train_cache.pkl at ./datasets/Holicity Load filename cache... [Info] Successfully initialized dataset Name: Holicity Mode: train Gt: None Counts: 12

[Info] Initializing Holicity dataset... Found filename cache holicity_test_cache.pkl at ./datasets/Holicity Load filename cache... [Info] Successfully initialized dataset Name: Holicity Mode: test Gt: None Counts: 3

     Successfully intialized dataloaders.

    --------Initializing model----------
     [Debug] Adding w_junc with value 0.000000 to model
     [Debug] Adding w_heatmap with value 0.000000 to model
    Model architecture: simple
    Backbone: lcnn
    Junction decoder: superpoint_decoder
    Heatmap decoder: pixel_shuffle
    -------------------------------------
     Successfully initialized model

================== Training ==================== Traceback (most recent call last): File "sold2/experiment.py", line 229, in main(args, dataset_cfg, model_cfg,export_dataset_mode=args.export_dataset_mode, device=device) File "sold2/experiment.py", line 112, in main train(args, dataset_cfg, model_cfg, output_path) File "sold2/experiment.py", line 82, in train train_net(args, dataset_cfg, model_cfg, output_path) File "/home/pc/SOLD2/sold2/train.py", line 179, in train_net epoch=epoch) File "/home/pc/SOLD2/sold2/train.py", line 253, in train_single_epoch junc_map = data["junction_map"].cuda() KeyError: 'junction_map'

[DennisJcy]

Fine,it's my mistake without uncommenting the gt source in the .yaml .

[rpautrat]

Glad that you solved it. I am closing this now.