CrazySssst
CrazySssst
do you solve this problem? I meet same issue when run RIDE agent(https://github.com/facebookresearch/impact-driven-exploration)
any one can send the pre-trained model to me ? i really need it, thanks for your help @linchaozhang @farmingyard @KeyKy @leolee96 @ananddb90
@ananddb90 this is my email : [email protected] , thank you !
thanks for you sharing. i will share my results on SSD after finish my experiment
@linchaozhang Do you pre-train the model in coco?
I met same problem, loss doesn't decrease.
forget to add expand_param in train_transform_param ? train_transform_param = { 'mirror': True, 'mean_value': [127.5, 127.5, 127.5], 'scale':0.007843, 'resize_param': { ... }, 'distort_param': { ... }, 'expand_param': { 'prob': 0.5, 'max_expand_ratio':...
why you set accum_batch_size=8 ? 8 is too small. batch_size = 8 accum_batch_size = 32 this setting work well for me
@chengyangfu @caiqi yes, i have the same confusion as you . in dssd branch , i find these code in prior_box_layer.cpp ``` for (int h = 0; h < layer_height;...
I meet the same issue. Do you solve it ? @sungwoong @tianjunz