fix(labs): bump mlsysim wheel ref 0.1.0 to 0.1.1 in all 33 labs
pyproject.toml was bumped to 0.1.1 in PR #1523 but the micropip.install() URLs in every lab still pointed to 0.1.0. This caused TestWheelConsistency and the WASM smoke test to fail on every subsequent PR. Bulk-replaced across all 33 labs in labs/vol1/ and labs/vol2/.
fix(ci): add .codespellrc to suppress false positive spell check failures
The codespell check has been failing on every PR because there was no ignore list. Added .codespellrc at the repo root:
Skips socratiq/src_shadow (vendored/compiled JS, not hand-written)
Whitelists: clos (Clos network topology), fpr (False Positive Rate), rin (ring buffer variable), ans, fo, curren (contributor name "Curren Iyer")
fix(staffml): correct tinyml-0384 KWS question
Option 3 distractor used a valid throughput approach (4 x 80 = 320 MFLOPS) then poisoned it with a false MHz=MFLOPS equivalence, which penalized students who reasoned correctly. Replaced with an unambiguously wrong distractor. Also:
Napkin math now shows both equivalent solution paths (latency and throughput)
common_mistake updated to flag the MHz vs MFLOPS confusion
fix(tinytorch): strip solution blocks when creating student notebooks
tito module start was creating notebooks from src/ verbatim via jupytext, including all working implementations between ### BEGIN SOLUTION / ### END SOLUTION markers. Students received solved notebooks instead of blank scaffolding.
_create_module_from_src now strips solution blocks and replaces them with raise NotImplementedError stubs before passing to jupytext. A _strip_solutions static method handles the stripping. Asset files alongside the source are copied to the temp dir so the notebook can still reference them.
uv run --with pytest python -m pytest tests/ in mlsysim/ -- 363 passed
_strip_solutions verified directly on src/01_tensor/01_tensor.py
TestWheelConsistency will pass: all 33 labs now reference mlsysim-0.1.1
Codespell will pass: false positives suppressed via .codespellrc
🔄 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/1542
**Author:** [@Shashank-Tripathi-07](https://github.com/Shashank-Tripathi-07)
**Created:** 4/25/2026
**Status:** ❌ Closed
**Base:** `dev` ← **Head:** `fix/close-issues-1531-1532-1502-1508`
---
### 📝 Commits (1)
- [`158a2cf`](https://github.com/harvard-edge/cs249r_book/commit/158a2cfd4563cc0b6426fe426a8e5b2f9f7edefb) fix: close issues 1531, 1532, 1502, 1508
### 📊 Changes
**36 files changed** (+96 additions, -48 deletions)
<details>
<summary>View changed files</summary>
➕ `.codespellrc` (+3 -0)
📝 `interviews/vault/questions/tinyml/tinyml-0384.yaml` (+14 -13)
📝 `labs/vol1/lab_00_introduction.py` (+1 -1)
📝 `labs/vol1/lab_01_ml_intro.py` (+1 -1)
📝 `labs/vol1/lab_02_ml_systems.py` (+1 -1)
📝 `labs/vol1/lab_03_ml_workflow.py` (+1 -1)
📝 `labs/vol1/lab_04_data_engr.py` (+1 -1)
📝 `labs/vol1/lab_05_nn_compute.py` (+1 -1)
📝 `labs/vol1/lab_06_nn_arch.py` (+1 -1)
📝 `labs/vol1/lab_07_ml_frameworks.py` (+1 -1)
📝 `labs/vol1/lab_08_model_train.py` (+1 -1)
📝 `labs/vol1/lab_09_data_selection.py` (+1 -1)
📝 `labs/vol1/lab_10_model_compress.py` (+1 -1)
📝 `labs/vol1/lab_11_hw_accel.py` (+1 -1)
📝 `labs/vol1/lab_12_perf_bench.py` (+1 -1)
📝 `labs/vol1/lab_13_model_serving.py` (+1 -1)
📝 `labs/vol1/lab_14_ml_ops.py` (+1 -1)
📝 `labs/vol1/lab_15_responsible_engr.py` (+1 -1)
📝 `labs/vol1/lab_16_ml_conclusion.py` (+1 -1)
📝 `labs/vol2/lab_01_introduction.py` (+1 -1)
_...and 16 more files_
</details>
### 📄 Description
Closes #1531, Closes #1532, Closes #1502, Closes #1508
## fix(labs): bump mlsysim wheel ref 0.1.0 to 0.1.1 in all 33 labs
`pyproject.toml` was bumped to `0.1.1` in PR #1523 but the `micropip.install()` URLs in every lab still pointed to `0.1.0`. This caused `TestWheelConsistency` and the WASM smoke test to fail on every subsequent PR. Bulk-replaced across all 33 labs in `labs/vol1/` and `labs/vol2/`.
## fix(ci): add .codespellrc to suppress false positive spell check failures
The codespell check has been failing on every PR because there was no ignore list. Added `.codespellrc` at the repo root:
- Skips `socratiq/src_shadow` (vendored/compiled JS, not hand-written)
- Whitelists: `clos` (Clos network topology), `fpr` (False Positive Rate), `rin` (ring buffer variable), `ans`, `fo`, `curren` (contributor name "Curren Iyer")
## fix(staffml): correct tinyml-0384 KWS question
Option 3 distractor used a valid throughput approach (4 x 80 = 320 MFLOPS) then poisoned it with a false MHz=MFLOPS equivalence, which penalized students who reasoned correctly. Replaced with an unambiguously wrong distractor. Also:
- Napkin math now shows both equivalent solution paths (latency and throughput)
- `common_mistake` updated to flag the MHz vs MFLOPS confusion
## fix(tinytorch): strip solution blocks when creating student notebooks
`tito module start` was creating notebooks from `src/` verbatim via jupytext, including all working implementations between `### BEGIN SOLUTION` / `### END SOLUTION` markers. Students received solved notebooks instead of blank scaffolding.
`_create_module_from_src` now strips solution blocks and replaces them with `raise NotImplementedError` stubs before passing to jupytext. A `_strip_solutions` static method handles the stripping. Asset files alongside the source are copied to the temp dir so the notebook can still reference them.
Verified on module 01: 13 solution blocks stripped, 13 stubs inserted, 0 markers remaining.
## Test plan
- [x] `uv run --with pytest python -m pytest tests/` in `mlsysim/` -- 363 passed
- [x] `_strip_solutions` verified directly on `src/01_tensor/01_tensor.py`
- [x] `TestWheelConsistency` will pass: all 33 labs now reference `mlsysim-0.1.1`
- [x] Codespell will pass: false positives suppressed via `.codespellrc`
---
<sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
📋 Pull Request Information
Original PR: https://github.com/harvard-edge/cs249r_book/pull/1542
Author: @Shashank-Tripathi-07
Created: 4/25/2026
Status: ❌ Closed
Base:
dev← Head:fix/close-issues-1531-1532-1502-1508📝 Commits (1)
158a2cffix: close issues 1531, 1532, 1502, 1508📊 Changes
36 files changed (+96 additions, -48 deletions)
View changed files
➕
.codespellrc(+3 -0)📝
interviews/vault/questions/tinyml/tinyml-0384.yaml(+14 -13)📝
labs/vol1/lab_00_introduction.py(+1 -1)📝
labs/vol1/lab_01_ml_intro.py(+1 -1)📝
labs/vol1/lab_02_ml_systems.py(+1 -1)📝
labs/vol1/lab_03_ml_workflow.py(+1 -1)📝
labs/vol1/lab_04_data_engr.py(+1 -1)📝
labs/vol1/lab_05_nn_compute.py(+1 -1)📝
labs/vol1/lab_06_nn_arch.py(+1 -1)📝
labs/vol1/lab_07_ml_frameworks.py(+1 -1)📝
labs/vol1/lab_08_model_train.py(+1 -1)📝
labs/vol1/lab_09_data_selection.py(+1 -1)📝
labs/vol1/lab_10_model_compress.py(+1 -1)📝
labs/vol1/lab_11_hw_accel.py(+1 -1)📝
labs/vol1/lab_12_perf_bench.py(+1 -1)📝
labs/vol1/lab_13_model_serving.py(+1 -1)📝
labs/vol1/lab_14_ml_ops.py(+1 -1)📝
labs/vol1/lab_15_responsible_engr.py(+1 -1)📝
labs/vol1/lab_16_ml_conclusion.py(+1 -1)📝
labs/vol2/lab_01_introduction.py(+1 -1)...and 16 more files
📄 Description
Closes #1531, Closes #1532, Closes #1502, Closes #1508
fix(labs): bump mlsysim wheel ref 0.1.0 to 0.1.1 in all 33 labs
pyproject.tomlwas bumped to0.1.1in PR #1523 but themicropip.install()URLs in every lab still pointed to0.1.0. This causedTestWheelConsistencyand the WASM smoke test to fail on every subsequent PR. Bulk-replaced across all 33 labs inlabs/vol1/andlabs/vol2/.fix(ci): add .codespellrc to suppress false positive spell check failures
The codespell check has been failing on every PR because there was no ignore list. Added
.codespellrcat the repo root:socratiq/src_shadow(vendored/compiled JS, not hand-written)clos(Clos network topology),fpr(False Positive Rate),rin(ring buffer variable),ans,fo,curren(contributor name "Curren Iyer")fix(staffml): correct tinyml-0384 KWS question
Option 3 distractor used a valid throughput approach (4 x 80 = 320 MFLOPS) then poisoned it with a false MHz=MFLOPS equivalence, which penalized students who reasoned correctly. Replaced with an unambiguously wrong distractor. Also:
common_mistakeupdated to flag the MHz vs MFLOPS confusionfix(tinytorch): strip solution blocks when creating student notebooks
tito module startwas creating notebooks fromsrc/verbatim via jupytext, including all working implementations between### BEGIN SOLUTION/### END SOLUTIONmarkers. Students received solved notebooks instead of blank scaffolding._create_module_from_srcnow strips solution blocks and replaces them withraise NotImplementedErrorstubs before passing to jupytext. A_strip_solutionsstatic method handles the stripping. Asset files alongside the source are copied to the temp dir so the notebook can still reference them.Verified on module 01: 13 solution blocks stripped, 13 stubs inserted, 0 markers remaining.
Test plan
uv run --with pytest python -m pytest tests/inmlsysim/-- 363 passed_strip_solutionsverified directly onsrc/01_tensor/01_tensor.pyTestWheelConsistencywill pass: all 33 labs now referencemlsysim-0.1.1.codespellrc🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.