rlworkgroup/garage

Add support for multiple layers for recurrent network models

Open

#709 opened on Jun 1, 2019

View on GitHub
 (0 comments) (0 reactions) (0 assignees)Python (1,728 stars) (297 forks)batch import
featuregood first issuetf

Description

Current recurrent network models, e.g. LSTMModel, only take single recurrent layer, which may not be sufficient for learning complex tasks. We should support multiple layers.

Contributor guide

Tech stack
python
Domain
machine learningbackend
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
mostly clear
Prerequisites
Python programmingRecurrent neural networksgarage codebase structure
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
40
Research direction
Examine the current implementation of LSTMModel and other recurrent models in the garage codebase (likely in garage/torch/models). Understand how the single recurrent layer is used. Add a 'num layers' parameter to the model constructor and modify the forward pass to stack multiple RNN layers. Ensure backward compatibility and add unit tests. Check for any existing discussions or PRs related to this issue.
Add support for multiple layers for recurrent network models · rlworkgroup/garage#709 | Good First Issue