what is the usage of buffer = vr.get_batch(index).asnumpy() when using decord to read frames? about ask-anything HOT 4 CLOSED

To get multiple frames at once, use get_batch this is the efficient way to obtain a long list of frames. Perhaps with decord.bridge.set_bridge('torch') could be a more efficient data pre-process method.

hi, but my question is not get_batch to obtains several frames meanwhile. My question is: buffer = vr.get_batch(index).asnumpy() the buffer tensor is never used. why use buffer before images_group?

from ask-anything.

To get multiple frames at once, use get_batch this is the efficient way to obtain a long list of frames. Perhaps with decord.bridge.set_bridge('torch') could be a more efficient data pre-process method.

from ask-anything.

Sorry for just realizing that this was a duplicate block that wasn't removed during code cleanup. It has been corrected, and indeed, that data block wasn't utilized.

from ask-anything.

thanks for your responsible work.

from ask-anything.