Coders In Hoods

Courses
Answers
Setup
Bookmarks

Coders In Hoods Academy

Every course is free. Sign in to start, and your progress, quiz scores and bookmarks follow you.

Academy

Courses
Paths
Exercise answers
Setup guides
Bookmarks

Coders In Hoods

The agency
Coaching for teams

Social

LinkedIn
X (Twitter)

Contact

codersinhoods@gmail.com

© Coders In Hoods · since 2019

Free · no card

Coders In Hoods Academy

Sign in to start learning

Free, no card. We use your account to save progress, quiz scores, bookmarks and occasionally send updates from our team.

Course overview

Fine-Tuning and RL for Agents

Free · sign in to start and track progress

1 · Decide and prepare

1Prompt, optimise or train?
2Training data from traces

2 · Supervised and preference fine-tuning

3SFT with LoRA on a free GPU
4Preference optimisation

3 · Reinforcement learning for agents

5GRPO in practice
6Designing rewards
Exercise answers
- Answer: Build and validate a composite triage reward
- Answer: Compare a relative judge with a pointwise one
7Multi-turn RL environments
8Reward hacking, regressions and shipping

CapstoneTrain a triage model, then decide whether to ship it Reference implementation

Answers Lesson 6: Designing rewards

Step-by-step answer

Core 75–90 min7 steps

Build and validate a composite triage reward

Back to the exercise

Previous answerReproduce TRL's loss aggregations by hand Next answer Compare a relative judge with a pointwise one

Free · sign in to start

Sign in to read the answer.

Free, no card. We use your account to save progress, quiz scores, bookmarks and occasionally send updates from our team.

7 more sections
Progress, quiz scores and bookmarks