浏览代码

modification to get started on the recurrent NN rewrite

/develop-generalizationTraining-TrainerController
vincentpierre 7 年前
当前提交
9933b56e
共有 1 个文件被更改,包括 6 次插入2 次删除
  1. 8
      unity-environment/Assets/ML-Agents/Scripts/CoreBrainInternal.cs

8
unity-environment/Assets/ML-Agents/Scripts/CoreBrainInternal.cs


runner.AddInput(graph[graphScope + ObservationPlaceholderName[obs_number]][0], observationMatrixList[obs_number]);
}
if (hasRecurrent)
{
runner.AddInput(graph[graphScope + RecurrentInPlaceholderName][0], inputOldMemories);
runner.Fetch(graph[graphScope + RecurrentOutPlaceholderName][0]);
}
TFTensor[] networkOutput;
try
{

{
Dictionary<int, float[]> new_memories = new Dictionary<int, float[]>();
runner.AddInput(graph[graphScope + RecurrentInPlaceholderName][0], inputOldMemories);
runner.Fetch(graph[graphScope + RecurrentOutPlaceholderName][0]);
float[,] recurrent_tensor = networkOutput[1].GetValue() as float[,];
int i = 0;

正在加载...
取消
保存