-
Notifications
You must be signed in to change notification settings - Fork 0
YXJiang1996/FormulaPrediction
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
1.潜在空间的维度,为什么是13 2.损失的权重 300,300,400 3.使用nice,rnvp,glow 4.使用dropout,不使用dropout 5.输入输出,padding,维度扩大 6.损失的权重 数据集说明: data_01.npz:21种色浆,选择3种生成配方,浓度0-1之间,使用km模型 模型说明: model_01: input+8*(GLOWCouplingBlock+PermuteRandom)+GLOWCouplingBlock+output n_epoch:3000 training_set:data_01 model_02: input+8*(GLOWCouplingBlock+PermuteRandom)+output n_epoch:1000 training_set:data_01 model_03: input+8*(RNVPCouplingBlock+PermuteRandom)+output n_epoch:1000 training_set:data_01 model_04: 测试结果说明: compare_01: model:model_01 test_set:data_01.npz 采样:256 比较:仅按色差大小排序进行配方选择时,同一个模型多次进行配方预测,最后选择的配方并不相同,结果也可能相差较大 问题:在神经网络输出一系列的配方后,如何进行最后预测配方的选择 compare_02: model:model_01 test_set:data_01.npz 采样:256,1024,4096 比较:采样点增多时,最后得到的结果有明显的改善 问题:采样点应该扩大到一个怎样的规模才算合适
About
No description, website, or topics provided.
Resources
Stars
Watchers
Forks
Releases
No releases published