Hello ! Thank you for your work
When downloading the test dataset on HF, i noticed an issue with the test_1 metadata: every video sequence shares the exact same w2c matrices. However, the videos themselves clearly show entirely different camera trajectories. It looks like a single trajectory might have been accidentally copied across all the test scenes during packaging.
will updated metadata files be available?
Thank you again !
Hello ! Thank you for your work
When downloading the test dataset on HF, i noticed an issue with the test_1 metadata: every video sequence shares the exact same w2c matrices. However, the videos themselves clearly show entirely different camera trajectories. It looks like a single trajectory might have been accidentally copied across all the test scenes during packaging.
will updated metadata files be available?
Thank you again !