Updates to track prediction (for demo_colmap.py)

[tommybotch]

Currently, predict_tracks_in_chunks doesn't detach tensors from GPU on each loop (see here).

I think this is partially causing related memory issues with demo_colmap.py as reported in #238.

In addition, memory load seems to be most affected by max_points_num (dropping this number increases chunking and lowers memory load).

[meta-cla]

Hi @tommybotch!

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!

[meta-cla]

Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Meta Open Source project. Thanks!

[1e100]

FWIW, this doesn't seem to solve the OOMs for me. GPU: A6000 (Ampere). Is there a reason to .detach() when the tensors were already created under the no_grad() context manager and do not require gradients?