Getting the performance w.r.t. different sparsity levels.
theurn15 opened this issue · 0 comments
theurn15 commented
Dear Author,
With regard to Getting the performance w.r.t. different sparsity levels, I tried to use the code provided by you (Tensorflow version) with different parameters, for instance:
python NGCF.py --dataset gowalla --regs [1e-5] --embed_size 64 --layer_size [64,64,64] --lr 0.0005 --pretrain 0 --batch_size 1024 --epoch 1000 --verbose 5 --node_dropout [0.1] --mess_dropout [0.1,0.1,0.1] --test_flag 'full' --report 1 --keep_prob [0.9,0.9]
However, I always get lower results than the one yoe reported in "Figure 4: Performance comparison over the sparsity..."
Hence, would you please release examples of the parameters string to Get the performance w.r.t. different sparsity levels?
Regards