Open eugenelawrence opened 4 years ago
Any Update? It seems the released code cannot reproduce the 73.12 miou on Pascal dataset.
@eugenelawrence are you able to run and train this model ?
@eugenelawrence are you able to run and train this model ?
No, the authors do not provide the pre-trained model. Thus I cannot run this model. And I have tried to re-implement the model based on this model, but I failed to reproduce the performance mentioned in the paper.
@hlzhu09 Hi,Hailong, could you please release the trained weight ASAP? I told you so many times
Any Update? It seems the released code cannot reproduce the 73.12 miou on Pascal dataset.
During training, we use some tricks in semantic segmentation and spend a lot of time finetuning some hyper parameters. I plan to release the training code. But the student has graduated and I am quite not sure when he can make it done …: (
Any Update? It seems the released code cannot reproduce the 73.12 miou on Pascal dataset.
During training, we use some tricks in semantic segmentation and spend a lot of time finetuning some hyper parameters. I plan to release the training code. But the student has graduated and I am quite not sure when he can make it done …: (
Thanks! Hope to see the training code ASAP.
When will you release the code?