TencentARC / ST-LLM

[ECCV 2024🔥] Official implementation of the paper "ST-LLM: Large Language Models Are Effective Temporal Learners"
Apache License 2.0
80 stars 2 forks source link

Questions about the testing strategy #1

Closed Richar-Du closed 3 months ago

Richar-Du commented 3 months ago

Thanks for your awesome work! I wonder whether the video is masked during testing?

farewellthree commented 3 months ago

Thank you for your attention. Masks were not used during the testing. We employed a frame sampling strategy with an FPS of 1 and a maximum of 16 frames.

Richar-Du commented 3 months ago

Got it. Thanks for your reply :)