I try to reproduce your code,but it seems that your file named utils.py lacks a function called "wrap" about gla-gcn HOT 6 OPEN

Hi!
I found wrap function in this repo and it works just fine!
https://github.com/vvirgooo2/MPM/blob/c0c58a572735057d6fbb1a06e4606c41bcd17b3e/common/camera.py#L23

from gla-gcn.

No description provided.
Hi there, thank you for the interest. This code is for an earlier version.
Need time to upload the latest code together with the fine-tune module. Please keep tuned.

from gla-gcn.

OK,thanks

from gla-gcn.

Hi, @skuley

I have encountered the following error message when using it, have you experienced the same issue?

python3 run_s-agcn.py -tta --evaluate 96_gt_243_supervised.bin
Namespace(dataset='h36m', keypoints='gt', subjects_train='S1,S5,S6,S7,S8', subjects_test='S9,S11', actions='*', checkpoint='checkpoint/ h36m', checkpoint_frequency=10, resume='', evaluate='96_gt_243_supervised.bin', render=False, by_subject=False, export_training_curves= True, stride=1, epochs=80, batch_size=512, dropout=0.1, learning_rate=0.01, lr_decay=0.95, data_augmentation=False, test_time_augmentation =True, architecture='3,3,3,3,3,3', causal=False, channels=96, subset=1, downsample=1, no_eval=False, dense=False, disable_optimizations= False, linear_projection=False, bone_length_term=True, no_proj=False, debug=False, viz_subject=None, viz_action=None, viz_camera=0, viz_ video=None, viz_skip=0, viz_output=None, viz_bitrate=3000, viz_no_ground_truth=False, viz_limit=-1, viz_downsample=1, viz_size=5)
Loading dataset...
Preparing data...
Loading 2D detections...
INFO: Receptive field: 243 frames
INFO: Trainable parameter count: 1348852
Loading checkpoint checkpoint/h36m/96_gt_243_supervised.bin
Traceback (most recent call last).
  File "/Users/rnic/Workspace/GLA-GCN/run_s-agcn.py", line 245, in <module>
    model_pos_train.load_state_dict(checkpoint['model_pos'])
  File "/opt/homebrew/lib/python3.11/site-packages/torch/nn/modules/module.py", line 2041, in load_state_dict
    raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for S_AGCN.
	Missing key(s) in state_dict: "data_bn.weight", "data_bn.bias", "data_bn.running_mean", "data_bn.running_var", "expand_gcn.gcn1.PA", " expand_gcn.gcn1.conv_a.0.weight", "expand_gcn.gcn1.conv_a.0.bias", "expand_gcn.gcn1.conv_a.1.weight", "expand_gcn.gcn1.conv_a.1.bias" , "expand_gcn.gcn1.conv_a.2.weight", "expand_gcn.gcn1.conv_a.2.bias", "expand_gcn.gcn1.conv_b.0.weight", "expand_gcn.gcn1.conv_b.0. bias", "expand_gcn.gcn1.conv_b.1.weight", "expand_gcn.gcn1.conv_b.1.bias", "expand_gcn.gcn1.conv_b.2.weight", "expand_gcn.gcn1.conv_b. b.2.bias", "expand_gcn.gcn1.conv_d.0.weight", "expand_gcn.gcn1.conv_d.0.bias", "expand_gcn.gcn1.conv_d.1.weight", "expand_gcn.gcn1. conv_d.1.bias", "expand_gcn.gcn1.conv_d.2.weight", "expand_gcn.gcn1.conv_d.2.bias", "expand_gcn.gcn1.down.0.weight", "expand_gcn.gcn1. down.0.bias", "expand_gcn.gcn1.down.1.weight", "expand_gcn.gcn1.down.1.bias", "expand_gcn.gcn1.down.1.running_mean", "expand_gcn.gcn1. down.1.running_var", ...

from gla-gcn.

Hi, @skuley

I have encountered the following error message when using it, have you experienced the same issue?

python3 run_s-agcn.py -tta --evaluate 96_gt_243_supervised.bin
Namespace(dataset='h36m', keypoints='gt', subjects_train='S1,S5,S6,S7,S8', subjects_test='S9,S11', actions='*', checkpoint='checkpoint/ h36m', checkpoint_frequency=10, resume='', evaluate='96_gt_243_supervised.bin', render=False, by_subject=False, export_training_curves= True, stride=1, epochs=80, batch_size=512, dropout=0.1, learning_rate=0.01, lr_decay=0.95, data_augmentation=False, test_time_augmentation =True, architecture='3,3,3,3,3,3', causal=False, channels=96, subset=1, downsample=1, no_eval=False, dense=False, disable_optimizations= False, linear_projection=False, bone_length_term=True, no_proj=False, debug=False, viz_subject=None, viz_action=None, viz_camera=0, viz_ video=None, viz_skip=0, viz_output=None, viz_bitrate=3000, viz_no_ground_truth=False, viz_limit=-1, viz_downsample=1, viz_size=5)
Loading dataset...
Preparing data...
Loading 2D detections...
INFO: Receptive field: 243 frames
INFO: Trainable parameter count: 1348852
Loading checkpoint checkpoint/h36m/96_gt_243_supervised.bin
Traceback (most recent call last).
  File "/Users/rnic/Workspace/GLA-GCN/run_s-agcn.py", line 245, in <module>
    model_pos_train.load_state_dict(checkpoint['model_pos'])
  File "/opt/homebrew/lib/python3.11/site-packages/torch/nn/modules/module.py", line 2041, in load_state_dict
    raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for S_AGCN.
	Missing key(s) in state_dict: "data_bn.weight", "data_bn.bias", "data_bn.running_mean", "data_bn.running_var", "expand_gcn.gcn1.PA", " expand_gcn.gcn1.conv_a.0.weight", "expand_gcn.gcn1.conv_a.0.bias", "expand_gcn.gcn1.conv_a.1.weight", "expand_gcn.gcn1.conv_a.1.bias" , "expand_gcn.gcn1.conv_a.2.weight", "expand_gcn.gcn1.conv_a.2.bias", "expand_gcn.gcn1.conv_b.0.weight", "expand_gcn.gcn1.conv_b.0. bias", "expand_gcn.gcn1.conv_b.1.weight", "expand_gcn.gcn1.conv_b.1.bias", "expand_gcn.gcn1.conv_b.2.weight", "expand_gcn.gcn1.conv_b. b.2.bias", "expand_gcn.gcn1.conv_d.0.weight", "expand_gcn.gcn1.conv_d.0.bias", "expand_gcn.gcn1.conv_d.1.weight", "expand_gcn.gcn1. conv_d.1.bias", "expand_gcn.gcn1.conv_d.2.weight", "expand_gcn.gcn1.conv_d.2.bias", "expand_gcn.gcn1.down.0.weight", "expand_gcn.gcn1. down.0.bias", "expand_gcn.gcn1.down.1.weight", "expand_gcn.gcn1.down.1.bias", "expand_gcn.gcn1.down.1.running_mean", "expand_gcn.gcn1. down.1.running_var", ...

I have the same problem.

from gla-gcn.

@rniczh @faajabbari
this is just saying that model weights are missing such keywords. In other words, there are no weights to those keyword layers.

I think i trained on scratch.

Ill update if I did other way, its been a while I've looked into this repo :(

from gla-gcn.