CarperAI/trlx

Collapse reference+learner hydra heads when using LoRa

Open

#320 opened on Feb 21, 2023

View on GitHub
 (6 comments) (2 reactions) (0 assignees)Python (450 forks)batch import
contributions welcomefeature requestgood first issue

Repository metrics

Stars
 (4,184 stars)
PR merge metrics
 (No merged PRs in 30d)

Description

🚀 The feature, motivation, and pitch

With additive (delta-style) parameter-efficient tuning methods such as LoRa, we should be able to make a slightly more mem-efficient hydra architecture by using a single block that does ~frozen_head + tunable_weights for the learner/policy head's fwd-pass and simply frozen_head for the reference, instead of maintaining 2x heads.

CC @LouisCastricato and @cat-state for pointing this out

Alternatives

No response

Additional context

No response

Contributor guide

Research direction
Study the current hydra head implementation in the codebase. Understand how LoRa works and refactor the heads to use a single block that applies a frozen head for the reference and adds a tunable LoRa layer for the learner head.
Tech stack
python
Domain
machine learningai
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.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
PythonGit
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
65