mirror of
https://github.com/MLSysBook/TinyTorch.git
synced 2026-06-01 07:35:58 -05:00
777b0f1ce1
🔄 Changes: - Removed modules/source/08_optimizers/tests/ directory - Updated module.yaml to reference inline tests - All testing now handled within optimizers_dev.py file - Cleaned up pytest cache references ✅ Verification: - All inline tests still pass correctly - SGD and Adam optimizers working perfectly - Training integration demonstrating convergence - Module fully functional with inline testing approach This aligns with the decision to drop separate test files and rely on inline testing within the _dev.py files for immediate feedback and validation.
61 lines
1.5 KiB
YAML
61 lines
1.5 KiB
YAML
name: "optimizers"
|
|
title: "Optimizers - Gradient-Based Parameter Updates"
|
|
description: "Build intelligent optimization algorithms that enable effective neural network training"
|
|
version: "1.0.0"
|
|
author: "TinyTorch Team"
|
|
|
|
# Learning objectives
|
|
learning_objectives:
|
|
- "Understand gradient descent and how optimizers use gradients to update parameters"
|
|
- "Implement SGD with momentum for accelerated convergence"
|
|
- "Build Adam optimizer with adaptive learning rates for modern deep learning"
|
|
- "Master learning rate scheduling strategies for training stability"
|
|
- "See how optimizers enable complete neural network training workflows"
|
|
|
|
# Prerequisites
|
|
prerequisites:
|
|
- "01_tensor"
|
|
- "07_autograd"
|
|
|
|
# Module metadata
|
|
metadata:
|
|
difficulty: "expert"
|
|
time_estimate: "6-8 hours"
|
|
pedagogical_framework: "Build → Use → Analyze"
|
|
|
|
# Key concepts covered
|
|
concepts:
|
|
- "Gradient descent theory"
|
|
- "SGD with momentum"
|
|
- "Adam optimizer"
|
|
- "Learning rate scheduling"
|
|
- "Training loop integration"
|
|
|
|
# Exports to tinytorch package
|
|
exports:
|
|
- "SGD"
|
|
- "Adam"
|
|
- "StepLR"
|
|
- "gradient_descent_step"
|
|
|
|
# Files in this module
|
|
files:
|
|
main: "optimizers_dev.py"
|
|
readme: "README.md"
|
|
# tests: inline in optimizers_dev.py
|
|
|
|
# Assessment configuration
|
|
assessment:
|
|
total_points: 70
|
|
breakdown:
|
|
gradient_descent: 10
|
|
sgd_optimizer: 15
|
|
adam_optimizer: 20
|
|
scheduler: 10
|
|
training_integration: 15
|
|
|
|
# Next steps
|
|
next_modules:
|
|
- "09_training"
|
|
- "10_compression"
|
|
- "13_mlops" |