Has anyone successfully reproduced the results?

[ori0n112]

I see that many issues have problems that cannot be reproduced. Has this been resolved now?

[McGregorWwww]

First of all, many thanks to the original authors for sharing such an inspiring idea. However, it seems that reproducibility has been a common challenge.

If you're looking for a highly reproducible implementation along similar lines, you might be interested in our ICCV 2025 work NEURONS, which demonstrates consistent and solid improvements over NeuroClips. We provide complete code, pretrained weights, training logs, configuration files, and generated videos, all publicly available to ensure full reproducibility and support further research.

[showlibia]

First of all, many thanks to the original authors for sharing such an inspiring idea. However, it seems that reproducibility has been a common challenge.

If you're looking for a highly reproducible implementation along similar lines, you might be interested in our ICCV 2025 work NEURONS, which demonstrates consistent and solid improvements over NeuroClips. We provide complete code, pretrained weights, training logs, configuration files, and generated videos, all publicly available to ensure full reproducibility and support further research.

I find that your paper reports results that are different from, but close to, the original NeuroClips paper. How did you reproduce the experiments? I am having problems with reproduction.