Hello author, can you publish the parameters of EPIC dataset? I trained the model according to the parameters in the paper and the parameters published on github, the mAP were only 24.5 and 22.24.Looking forward to your reply, thanks a lot, it is very important to me.
We have released the weights in the README (see here). We train the model using the same parameters as the released code. From some recent works, it seems that different GPUs can lead to different results for methods with the same codebase.
dingfengshi
commented
9 months ago
Hello author, can you publish the parameters of EPIC dataset? I trained the model according to the parameters in the paper and the parameters published on github, the mAP were only 24.5 and 22.24.Looking forward to your reply, thanks a lot, it is very important to me.