ikostrikov / ikostrikov/implicit_q_learning

Why use a positive learning rate in finetuning?

Open
#10 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
337
Forks
49
PR merge metrics
No merged PRs in 30d

Description

In the file train_finetune, this code schedule_fn = optax.cosine_decay_schedule(-actor_lr, max_steps) seems to use a positive learning rate?why?

Contributor guide

No contributing guide indexed for this repository

Research direction

Start with the train_finetune file and inspect the schedule_fn assignment using optax.cosine_decay_schedule. Clarify in the issue or relevant documentation why the learning-rate value is passed with its shown sign, and consider the question resolved when that behavior is explicitly explained.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Documentation
Difficulty
1/5
Estimated time
Under an hour
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.