Confusion about PyTorch checkpoint for online BootsTAPIR

[weirenorweiren]

I realize the PyTorch checkpoints for online BootsTAPIR in torch_causal_tapir_demo.ipynb and README are not the same. The former is https://storage.googleapis.com/dm-tapnet/causal_bootstapir_checkpoint.pt but the later is https://storage.googleapis.com/dm-tapnet/bootstap/causal_bootstapir_checkpoint.pt with an extra directory of bootstap. The one for the demo works okay but there are runtime problems for the checkpoint in README. I have listed all problems below. Could you please clarify on the following questions?

1. Which checkpoint is the right one to use? Now only the one in demo works but I am not sure whether that's the right one.
2. If the one in README is the right one, how to make it work?
3. Do you have a PyTorch checkpoint for online TAPIR?

pyramid_level=1

wget -P F:\Wei\tapnet\tapnet\checkpoints https://storage.googleapis.com/dm-tapnet/bootstap/causal_bootstapir_checkpoint.pt
model = tapir_model.TAPIR(pyramid_level=1, use_casual_conv=True)
model.load_state_dict(torch.load("tapnet/checkpoints/causal_bootstapir_checkpoint.pt"))

Traceback (most recent call last):
  File "F:\Wei\tapnet\tapnet\live_demo_thorlab_camera_torch.py", line 17, in <module>
    model.load_state_dict(torch.load("tapnet/checkpoints/causal_bootstapir_checkpoint.pt"))
  File "C:\Users\NOCB\anaconda3\envs\tapnet_torch\Lib\site-packages\torch\nn\modules\module.py", line 2215, in load_state_dict
    raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for TAPIR:
        size mismatch for extra_convs.blocks.0.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).
        size mismatch for extra_convs.blocks.1.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).
        size mismatch for extra_convs.blocks.2.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).
        size mismatch for extra_convs.blocks.3.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).
        size mismatch for extra_convs.blocks.4.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).

pyramid_level=0

wget -P F:\Wei\tapnet\tapnet\checkpoints https://storage.googleapis.com/dm-tapnet/bootstap/causal_bootstapir_checkpoint.pt
model = tapir_model.TAPIR(pyramid_level=0, use_casual_conv=True)
model.load_state_dict(torch.load("tapnet/checkpoints/causal_bootstapir_checkpoint.pt"))

Traceback (most recent call last):
  File "F:\Wei\tapnet\tapnet\live_demo_thorlab_camera_torch.py", line 17, in <module>
    model.load_state_dict(torch.load("tapnet/checkpoints/causal_bootstapir_checkpoint.pt"))
  File "C:\Users\NOCB\anaconda3\envs\tapnet_torch\Lib\site-packages\torch\nn\modules\module.py", line 2215, in load_state_dict
    raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for TAPIR:
        size mismatch for torch_pips_mixer.linear.weight: copying a param with shape torch.Size([512, 535]) from checkpoint, the shape in current model is torch.Size([512, 486]).
        size mismatch for extra_convs.blocks.0.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).
        size mismatch for extra_convs.blocks.1.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).
        size mismatch for extra_convs.blocks.2.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).
        size mismatch for extra_convs.blocks.3.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).
        size mismatch for extra_convs.blocks.4.conv.weight: copying a param with shape torch.Size([1024, 768, 3, 3]) from checkpoint, the shape in current model is torch.Size([1024, 256, 3, 3]).