I am currently working with the Argoverse dataset, leveraging a pre-trained inference model. While I am not directly utilizing the official Argoverse repository, I am loading Argoverse data for inference purposes through this external model.
However, I have encountered a challenge with the calibration matrix, which appears to be a required component for the model’s processing pipeline. Unfortunately, I do not have access to the calibration matrices, nor do I have clarity on how to compute or retrieve them independently from the dataset.
I would greatly appreciate any guidance or recommendations on how to obtain or generate the appropriate calibration matrices for Argoverse data. Specifically:
Are these matrices available in a standard format within the original dataset?
Is there a documented method or script for calculating them if they are missing?
Would there be a way to adapt or bypass this step when using pre-trained models?
Any insights or references would be incredibly helpful.
Thank you in advance for your support.
I am currently working with the Argoverse dataset, leveraging a pre-trained inference model. While I am not directly utilizing the official Argoverse repository, I am loading Argoverse data for inference purposes through this external model.
However, I have encountered a challenge with the calibration matrix, which appears to be a required component for the model’s processing pipeline. Unfortunately, I do not have access to the calibration matrices, nor do I have clarity on how to compute or retrieve them independently from the dataset.
I would greatly appreciate any guidance or recommendations on how to obtain or generate the appropriate calibration matrices for Argoverse data. Specifically:
Are these matrices available in a standard format within the original dataset?
Is there a documented method or script for calculating them if they are missing?
Would there be a way to adapt or bypass this step when using pre-trained models?
Any insights or references would be incredibly helpful.
Thank you in advance for your support.