your results presented at here is not the same in your paper about video_captioning_rl HOT 2 CLOSED

If you see the bold note mentioned in the readme below the results table, it clearly says the reasons why the numbers are different from the paper. I am pasting those again here below:

Note that this released baseline model has several improvements (and significantly better results) w.r.t. our EMNLP17 paper results, due to the following enhancements:

Better visual features: (1) ResNet-152 frame-level features (2) ResNeXt-101 motion features
Used SCST approach instead of MIXER for reinforcement leanring
Used better entailment classifier

from video_captioning_rl.

Thank you, I see it.

from video_captioning_rl.