Error running eval for all frames - samples mismatch

[jamalknight]

Hi there

I got an error during eval (triangulation) when running the h36m dataset. The only thing changed is that I pre processed every frame instead of every 64th frame. I want to see what it looks like on sequential frames instead of skipping.

Command:
python3 train.py --eval --eval_dataset val --config experiments/human36m/eval/human36m_vol_softmax.yaml --logdir ./logs

Error:
AssertionError: [train=True, test=False] ./data/human36m/extra/human36m-multiview-labels-GTbboxes.npy has 389938 samples, but './data/pretrained/human36m/human36m_alg_10-04-2019/checkpoints/0060/results/train.pkl' has 159181. Did you follow all preprocessing instructions carefully?

How do I fix the 'samples' mismatch?

[jamalknight]

Needed to recalculate pelvis positions for all frames