rlworkgroup/garage

Split GaussianMLPModel into several variants

開放

#587 建立於 2019年3月19日

 (0 則留言) (0 個反應) (0 位負責人)Python (297 個分叉)batch import
good first issuetf

倉庫指標

星標
 (1,728 顆星)
PR 合併指標
 (30 天內沒有已合併 PR)

描述

The current GaussianMLPModel actually represents several architecture variants, which makes it difficult to use and to test. We should split it into several variants based on architecture (and possibly parameterization)

description mean std_param
fixed variance MLP1 parameter (trainable or not)
split network MLP1 MLP2
shared network/multi-headed MLP1/head1 MLP1/head2

貢獻者指南