- Add git pull --rebase before push so concurrent comment-triggered runs
don't reject each other (only one of three runs had succeeded for PR 1179).
- Manually add @salmanmkc as code contributor to tinytorch and kits
(labs was already added by the single successful run).
- Fix Pareto diagram: swap A/C so line has positive slope (latency vs accuracy)
- Add missing Image Classification writeup to Standard Benchmark Tasks
- Treat anomaly_detection as binary classification (np.rand(2))
- Convert MLPerf inputs to Tensors + transpose HWC→CHW for TinyTorch models
Fixesharvard-edge/cs249r_book#1196
The GitHub link (github.com/harvard-edge/TinyTorch/blob/main/paper/paper.pdf)
returns 404. Use arxiv.org/pdf/2601.19107 instead.
Fixesharvard-edge/cs249r_book#1198
Add encoding='utf-8' and errors='replace' to subprocess.run() calls in
workflow.py so unit and integration test output decode correctly on
Windows (cp1252) when output contains UTF-8 characters.
Co-authored-by: Pratham-ja <114498234+Pratham-ja@users.noreply.github.com>
Refines the explanation of K,V computation savings in the memoization module,
quantifying redundant computations and highlighting the efficiency gain.
The paper and module now specify that generating 100 tokens requires 5,050
total K,V computations, but only 100 are necessary, resulting in 4,950
redundant calculations.
ngbolin was correctly added to tinytorch (PR #1180) but the edited-comment
re-trigger on PR #1181 ran the old LLM code which hallucinated ngbolin as
the username instead of pipme.
The LLM (llama3.1:8b) was hallucinating usernames — e.g. returning
"pipmea" instead of "pipme". Since the username is always present as
an @mention in the trigger line, extract it deterministically via regex
in Step 1 and only use the LLM to classify contribution types.
Notebooks use short names (tensor.ipynb, not 01_tensor.ipynb) but docs
and Binder postBuild scripts used the prefixed form. This caused broken
Binder links and incorrect paths in troubleshooting guides.
Fixes: harvard-edge/cs249r_book#1176
Replace hardcoded numerical values across all module ABOUT.md files with
Python-computed values using myst_nb glue() references. Each file is now a
MyST Markdown Notebook that executes inline code cells to compute memory
sizes, FLOPs, compression ratios, and other quantitative values.
Key changes:
- Add file_format: mystnb frontmatter and code-cell blocks to all 20 files
- All arithmetic (memory calculations, speedups, ratios) now computed inline
- Fix multiple arithmetic errors discovered during conversion
- Enable execute_notebooks: "cache" in PDF config for glue resolution
- Fix jupyter-book version constraint in Makefile
Mermaid diagrams were oversized in PDF output. Reduced viewport width
from 800→600 and added LaTeX preamble to cap mermaid figures at
0.75\linewidth. Also fixed 7 admonition blocks across 5 ABOUT.md files
where nested triple-backtick code fences broke the MyST parser, causing
raw markdown to render in PDF output.
- Responsive progressive disclosure: hide date at 1024px, icons-only at 768px
- Add 44px touch targets for mobile nav links
- Restyle version badge with separate number/date spans
- Add link-secondary class for visual hierarchy in navbar
- Add npx pre-flight check for Mermaid CLI in PDF build
- Configure mermaid_cmd to use npx (no global install needed)
- Add Unicode box-drawing fallbacks for LaTeX
- Add Community section to PDF table of contents
Use vscode.openWith(uri, 'jupyter-notebook') for .ipynb files so they
render in VS Code's notebook editor with cell support, rather than
opening as raw JSON text via openTextDocument.
