I appreciate your interest in our work and your desire to utilize our hash encoder. I have provided one checkpoint in the checkpoints/ folder. However, I'd like to bring to your attention a couple of considerations that may affect the feasibility of sharing the encoder checkpoint:
The encoder checkpoint I have may not necessarily align with the current state of the codebase available on GitHub. Given the time since the experiments, there may have been updates, bug fixes, or modifications to the code that could render the pre-trained encoder incompatible or less effective when used with the current codebase.
While I have provided the pre-trained encoder, it's important to note that even with the encoder in hand, its effective use will still require data preprocessing. Self-supervised training typically involves substantial data preprocessing steps essential for achieving optimal performance. As such, obtaining the encoder alone may not be sufficient for immediate use. I would strongly recommend that you consider training the encoder yourself with your specific dataset and preprocessing pipeline to ensure optimal results.
I appreciate your interest in our work and your desire to utilize our hash encoder. I have provided one checkpoint in the checkpoints/ folder. However, I'd like to bring to your attention a couple of considerations that may affect the feasibility of sharing the encoder checkpoint:
The encoder checkpoint I have may not necessarily align with the current state of the codebase available on GitHub. Given the time since the experiments, there may have been updates, bug fixes, or modifications to the code that could render the pre-trained encoder incompatible or less effective when used with the current codebase.
While I have provided the pre-trained encoder, it's important to note that even with the encoder in hand, its effective use will still require data preprocessing. Self-supervised training typically involves substantial data preprocessing steps essential for achieving optimal performance. As such, obtaining the encoder alone may not be sufficient for immediate use. I would strongly recommend that you consider training the encoder yourself with your specific dataset and preprocessing pipeline to ensure optimal results.