Fix float32 auto-conversion

[murlock1000]

Fixes #39

On RTX 4090 GPU the model code auto-converted aggregated and patch tokens to torch.float32 dtype, even though bfloat16/float16 was specified. This lead to twice the amount of memory being used on the GPU. This fix prevents the auto-conversion providing memory performance as stated in the paper.

[facebook-github-bot]

Hi @murlock1000!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks!

[perone]

Just a heads-up that the PR is showing the following below when using demo_colmap.py:

  File "/root/vggt/demo_colmap.py", line 298, in <module>
    demo_fn(args)
  File "/root/vggt/demo_colmap.py", line 139, in demo_fn
    extrinsic, intrinsic, depth_map, depth_conf = run_VGGT(model, images, dtype, vggt_fixed_resolution)
                                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/vggt/demo_colmap.py", line 86, in run_VGGT
    extrinsic = extrinsic.squeeze(0).cpu().numpy()
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: Got unsupported ScalarType BFloat16

So you have to fix the demo_colmap.py to convert to fp32 or fp16 before getting the extrinsics to numpy.