shyamal-b / ss-tad

End-to-End, Single-Stream Temporal Action Detection in Untrimmed Videos (Official Repo for SS-TAD)
107 stars 22 forks source link

When will the final version be released? #4

Open JonghwanMun opened 6 years ago

JonghwanMun commented 6 years ago

I wonder when the final version of code (e.g.training code, demo notebook) is available.

lsy323 commented 6 years ago

It has been over 10 months since the preview version of code released... Recently I implemented SS-TAD based on the released SST codes in Pytorch (unofficial repo but from a member of Stanford Vision Lab), but I got very poor results. I have no idea whether there is something wrong with my implementation or the selection of hyperparameters. (quite a lot of hyperparameters in SS-TAD model) I really hope the complete code and trained model will be released.

dmenig commented 5 years ago

Up !

dmenig commented 5 years ago

Seems like one of the only clean versions of ActivityNet end to end training (provided that the training script is released !)