Dongyuxin
Dongyuxin
@chengshuai , your layer { bottom: "conv_new_1" this is wrong connection , change it to bottom: "res5c" ################## layer { name: "bbox_pred_voc" …… inner_product_param { num_output: 8 why your output...
####-----------------conv_new2 layer { bottom: "conv_new_1" top: "conv_new_2" name: "conv_new_2" type: "Convolution" convolution_param { num_output: 490 /---------------------------/ here conv_new2's bottom is bottom: "res5c", so the paper's section. you can put your...
@Tete-Xiao ,i try hard to download dataset and config env, but sh demo_test.sh Namespace(arch_decoder='ppm_bilinear_deepsup', arch_encoder='resnet50_dilated8', batch_size=1, fc_dim=2048, gpu_id=0, imgMaxSize=1000, imgSize=[300, 400, 500, 600], model_path='baseline-resnet50_dilated8-ppm_bilinear_deepsup', num_class=150, num_val=-1, padding_constant=8, result='./', segm_downsampling_rate=8, suffix='_epoch_20.pth',...
can we collaborate on this Stereo lsd-slam code implement?
hi @mihaibujanca @amiltonwong @ho-pp @ustbguan , i come across some problem , https://github.com/tum-vision/lsd_slam/issues/203, this is very important and reference from https://www.ais.uni-bonn.de/theses/Nicola_Krombach_Master_Thesis_02_2016.pdf ,this give some advices on depth fusion. if anyone...