[PR #1114] [MERGED] fix: initialize parameter's gradient after creating Optimizer object #1124

New Issue

GiteaMirror · 2026-03-22T16:01:18-05:00

GiteaMirror commented

2026-03-22 16:01:18 -05:00

📋 Pull Request Information

Original PR: https://github.com/harvard-edge/cs249r_book/pull/1114
Author: @minhdang26403
Created: 1/19/2026
Status: ✅ Merged
Merged: 1/19/2026
Merged by: @profvjreddi

Base: dev ← Head: fix/optimizer-integration-test

📝 Commits (1)

8a8936f fix: initialize parameter's gradient after creating Optimizer object

📊 Changes

1 file changed (+16 additions, -15 deletions)

View changed files

📝 tinytorch/tests/07_optimizers/test_progressive_integration.py (+16 -15)

📄 Description

Currently, there are three tests that fail in the Optimizers module:

test_adam_optimizer_exists
test_sgd_optimizer_exists
test_zero_grad

The root cause of this issue is that the test code initializes the parameter's gradient before passing it to the Optimizer constructor, but the Optimizer's init method sets parameter's gradient to None. Therefore, we need to initialize the parameter's gradient after that.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/harvard-edge/cs249r_book/pull/1114 **Author:** [@minhdang26403](https://github.com/minhdang26403) **Created:** 1/19/2026 **Status:** ✅ Merged **Merged:** 1/19/2026 **Merged by:** [@profvjreddi](https://github.com/profvjreddi) **Base:** `dev` ← **Head:** `fix/optimizer-integration-test` --- ### 📝 Commits (1) - [`8a8936f`](https://github.com/harvard-edge/cs249r_book/commit/8a8936f22544979d31eb6d15dacb2f9806bf48eb) fix: initialize parameter's gradient after creating Optimizer object ### 📊 Changes **1 file changed** (+16 additions, -15 deletions) <details> <summary>View changed files</summary> 📝 `tinytorch/tests/07_optimizers/test_progressive_integration.py` (+16 -15) </details> ### 📄 Description Currently, there are three tests that fail in the Optimizers module: - test_adam_optimizer_exists - test_sgd_optimizer_exists - test_zero_grad The root cause of this issue is that the test code initializes the parameter's gradient before passing it to the Optimizer constructor, but the Optimizer's __init__ method sets parameter's gradient to None. Therefore, we need to initialize the parameter's gradient after that. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

GiteaMirror added the pull-request label 2026-03-22 16:01:18 -05:00

GiteaMirror closed this issue

2026-03-22 16:01:18 -05:00

GiteaMirror referenced this issue

2026-03-22 16:01:34 -05:00

[PR #1124] [CLOSED] Add TinyTorch-specific contributor recognition #1132

GiteaMirror referenced this issue

2026-04-11 08:25:10 -05:00

[PR #1124] [CLOSED] Add TinyTorch-specific contributor recognition #2355

GiteaMirror referenced this issue

2026-04-13 13:21:48 -05:00

[PR #1124] [CLOSED] Add TinyTorch-specific contributor recognition #3077