github-starred/cs249r_book
2
0
Fork 0
mirror of https://github.com/harvard-edge/cs249r_book.git synced 2026-07-15 21:28:33 -05:00
Code Issues 276 Packages Projects Releases 26 Wiki Activity

[PR #1869] [MERGED] fix(layers): Dropout uses global unseeded np.random instead of module rng #30969

New Issue
Closed
opened 2026-06-21 20:36:26 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-06-21 20:36:26 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/harvard-edge/cs249r_book/pull/1869
Author: @Shashank-Tripathi-07
Created: 6/16/2026
Status: ✅ Merged
Merged: 6/16/2026
Merged by: @profvjreddi

Base: main ← Head: fix/dropout-seeded-rng

📝 Commits (1)

  • 180e933 fix Dropout to use the module-level seeded rng instead of global np.random

📊 Changes

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

View changed files

📝 tinytorch/src/03_layers/03_layers.py (+1 -1)

📄 Description

What breaks

03_layers.py seeds a module-level RNG for reproducibility:

rng = np.random.default_rng(7)

Dropout._create_mask() ignores this and calls np.random.random(shape), which draws from the global unseeded state. Dropout masks are therefore different on every run even when the user expects deterministic behavior from the module seed. Any unit test that checks a specific Dropout output value will be flaky.

Fix

# before
binary_mask = (np.random.random(shape) < keep_prob).astype(np.float32)

# after
binary_mask = (rng.random(shape) < keep_prob).astype(np.float32)

One line change -- Dropout now participates in the same reproducible stream as every other random operation in the module.

Test plan

  • pytest tinytorch/tests/03_layers/ passes
  • Run twice with same seed -- Dropout produces identical masks both times

🔄 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/1869 **Author:** [@Shashank-Tripathi-07](https://github.com/Shashank-Tripathi-07) **Created:** 6/16/2026 **Status:** ✅ Merged **Merged:** 6/16/2026 **Merged by:** [@profvjreddi](https://github.com/profvjreddi) **Base:** `main` ← **Head:** `fix/dropout-seeded-rng` --- ### 📝 Commits (1) - [`180e933`](https://github.com/harvard-edge/cs249r_book/commit/180e9331971c039893c8bcd4b46347177866a2e3) fix Dropout to use the module-level seeded rng instead of global np.random ### 📊 Changes **1 file changed** (+1 additions, -1 deletions) <details> <summary>View changed files</summary> 📝 `tinytorch/src/03_layers/03_layers.py` (+1 -1) </details> ### 📄 Description ## What breaks `03_layers.py` seeds a module-level RNG for reproducibility: ```python rng = np.random.default_rng(7) ``` `Dropout._create_mask()` ignores this and calls `np.random.random(shape)`, which draws from the global unseeded state. Dropout masks are therefore different on every run even when the user expects deterministic behavior from the module seed. Any unit test that checks a specific Dropout output value will be flaky. ## Fix ```python # before binary_mask = (np.random.random(shape) < keep_prob).astype(np.float32) # after binary_mask = (rng.random(shape) < keep_prob).astype(np.float32) ``` One line change -- Dropout now participates in the same reproducible stream as every other random operation in the module. ## Test plan - [ ] `pytest tinytorch/tests/03_layers/` passes - [ ] Run twice with same seed -- Dropout produces identical masks both times --- <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 2026-06-21 20:36:26 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
area: book
area: collabs
area: kits
area: labs
area: socratiq
area: staffml
area: tinytorch
area: tools
area: website
bug
dependencies
format: epub
format: pdf
javascript
link-health
link-rot
priority-high
pull-request
Mirrored from GitHub Pull Request
 staffml
type: bug
type: citation
type: code
type: errata
type: improvement
type: new
type: question
vault-sli
No Label pull-request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/cs249r_book#30969
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?