Can we inference online video based on test_module? #14
Description
test_module may need n_segments=8, but can we inference with n_segments=1 and get same accuracy?
{{ message }}
test_module may need n_segments=8, but can we inference with n_segments=1 and get same accuracy?