scnn
scnn copied to clipboard
zhengshou
Segment-CNN: A Framework for Temporal Action Localization in Untrimmed Videos via Multi-stage CNNs
Dear zhengshou, When I ran the program, I found that the pretrained model URL has no content. Any help will be appreciated
Hi. I'm running the python demo file and i'm getting this error: >extract frames starts extract frames done in 10.043 s init sliding window starts init sliding window done in...
https://github.com/zhengshou/scnn/blob/ee05621eb272c80e72ce7a068f63e947fca3b112/demo/run_demo.py#L194 It appears that the loop should be `for item in range(row[2]-1, row[3])` instead of `for item in range(row[2]-1, row[3]+1)` since, at boundary cases (end of video file) it gives...
hey, as u showed in paper's sec4.3: results fine-tuning on cls network always perform better . So, why don't use it in prediction. If not, why we train it.
Hi Zhengshou, Thank you very much for your work on S-CNN. I'm currently trying to run the demo using C3D v1.1 but I realized the demo code was written for...
the outputs of proposal network (probabilities of proposals) are same and very low such as 2.48e-5. So, threshold 0.7 filters all proposals out.
when i run run_demo.py, the result show no such file or directory:"pred/pro/output/0.00001.prob I want to ask how to solve this problem. Many thanks！
Hi,zhengshou Thank you for your replying.I'm so sorry to trouble you again. I'm very interested in your work. So I have some questiones of how to prepare the training lst...
