The solutions were already implemented when I started module 1.
# Install TinyTorch (run from a project folder like ~/projects)
curl -sSL mlsysbook.ai/tinytorch/install.sh | bash
# Activate and verify
cd tinytorch
source .venv/bin/activate
tito setup
tito module start 01
Proposed Solution
Remove the implementations from the working notebooks.
Originally created by @bdub-1 on GitHub (May 5, 2026).
Original GitHub issue: https://github.com/harvard-edge/cs249r_book/issues/1684
### Module
01 Tensor
### Type of Improvement
Code quality (refactor, style, consistency)
### Description
The solutions were already implemented when I started module 1.
```
# Install TinyTorch (run from a project folder like ~/projects)
curl -sSL mlsysbook.ai/tinytorch/install.sh | bash
# Activate and verify
cd tinytorch
source .venv/bin/activate
tito setup
tito module start 01
```
### Proposed Solution
Remove the implementations from the working notebooks.
@Shashank-Tripathi-07 commented on GitHub (May 5, 2026):
The solutions you are seeing are in the source file (src/01_tensor/01_tensor.py), which is the contributor-facing source of truth. The ### BEGIN SOLUTION / ### END SOLUTION markers are intentional nbgrader annotations.
When a student runs tito src export, these blocks get stripped out to generate the student-facing notebook with empty stubs. The GitHub source view always shows the full implementation by design.
To get the blank student version locally:
cd cs249r_book/tinytorch
tito src export01
This generates the notebook in modules/01_tensor/ with only the TODO stubs visible. Not a bug, working as intended.
<!-- gh-comment-id:4380949453 -->
@Shashank-Tripathi-07 commented on GitHub (May 5, 2026):
The solutions you are seeing are in the **source file** (`src/01_tensor/01_tensor.py`), which is the contributor-facing source of truth. The `### BEGIN SOLUTION` / `### END SOLUTION` markers are intentional nbgrader annotations.
When a student runs `tito src export`, these blocks get **stripped out** to generate the student-facing notebook with empty stubs. The GitHub source view always shows the full implementation by design.
To get the blank student version locally:
```bash
cd cs249r_book/tinytorch
tito src export 01
```
This generates the notebook in `modules/01_tensor/` with only the TODO stubs visible. Not a bug, working as intended.
Thanks for the response. I used the quick start guide to being my module. I also don't see any mention of the this in the docs unless I missed it during the setup instructions. When I run tito --help there is no export command that shows up.
edit: I am following the student-learner role so I was confused why I am seeing the answers as my default. The TITO CLI module workflow page doesn't mention this either. I cannot get the blank version to generate so I am unsure if there actually is a bug. I saw a closed issue earlier of the same bug so I am wondering if the issue lingers
<!-- gh-comment-id:4381138846 -->
@bdub-1 commented on GitHub (May 5, 2026):
Thanks for the response. I used [the quick start guide](https://mlsysbook.ai/tinytorch/getting-started.html#start-the-module) to being my module. I also don't see any mention of the this in the docs unless I missed it during the setup instructions. When I run `tito --help` there is no `export` command that shows up.
edit: I am following the [student-learner role](https://mlsysbook.ai/tinytorch/tito/overview.html#commands-by-user-role) so I was confused why I am seeing the answers as my default. The [TITO CLI module workflow page](https://mlsysbook.ai/tinytorch/tito/modules.html) doesn't mention this either. I cannot get the blank version to generate so I am unsure if there actually is a bug. I saw a closed issue earlier of the same bug so I am wondering if the issue lingers
@Shashank-Tripathi-07 commented on GitHub (May 6, 2026):
This is a confirmed bug. When a student runs tito module start 01, the CLI was generating the notebook by passing the raw src/01_tensor/01_tensor.py file directly to jupytext with no stripping step. Every ### BEGIN SOLUTION / ### END SOLUTION block was carried over intact into the student notebook, so all answers were already filled in from the start.
The issue was in _create_module_from_src inside tinytorch/tito/commands/module/workflow.py. The function's own docstring acknowledged this behavior:
"Full src/ (including ### BEGIN SOLUTION ... ### END SOLUTION blocks) is passed through to jupytext"
The fix strips those solution blocks in memory before conversion. The source files in src/ are never touched. A temp directory is used to hold the stripped version, jupytext converts it to a notebook, and the temp dir is cleaned up automatically. Students now receive blank stubs with TODO comments, as intended.
<!-- gh-comment-id:4384835228 -->
@Shashank-Tripathi-07 commented on GitHub (May 6, 2026):
This is a confirmed bug. When a student runs `tito module start 01`, the CLI was generating the notebook by passing the raw `src/01_tensor/01_tensor.py` file directly to jupytext with no stripping step. Every `### BEGIN SOLUTION` / `### END SOLUTION` block was carried over intact into the student notebook, so all answers were already filled in from the start.
The issue was in `_create_module_from_src` inside `tinytorch/tito/commands/module/workflow.py`. The function's own docstring acknowledged this behavior:
> "Full `src/` (including `### BEGIN SOLUTION` ... `### END SOLUTION` blocks) is passed through to jupytext"
The fix strips those solution blocks in memory before conversion. The source files in `src/` are never touched. A temp directory is used to hold the stripped version, jupytext converts it to a notebook, and the temp dir is cleaned up automatically. Students now receive blank stubs with TODO comments, as intended.
PR: #1686
@Shashank-Tripathi-07 commented on GitHub (May 6, 2026):
Thanks for your attention to this bug @bdub-1 . I've made a PR and corrected this to make sure no future versions suffer from any problem like these !
We highly needed learner feedback for the project so that we can correct it further and better ! Thanks
<!-- gh-comment-id:4384840920 -->
@Shashank-Tripathi-07 commented on GitHub (May 6, 2026):
Thanks for your attention to this bug @bdub-1 . I've made a PR and corrected this to make sure no future versions suffer from any problem like these !
We highly needed learner feedback for the project so that we can correct it further and better ! Thanks
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.
Originally created by @bdub-1 on GitHub (May 5, 2026).
Original GitHub issue: https://github.com/harvard-edge/cs249r_book/issues/1684
Module
01 Tensor
Type of Improvement
Code quality (refactor, style, consistency)
Description
The solutions were already implemented when I started module 1.
Proposed Solution
Remove the implementations from the working notebooks.
@Shashank-Tripathi-07 commented on GitHub (May 5, 2026):
The solutions you are seeing are in the source file (
src/01_tensor/01_tensor.py), which is the contributor-facing source of truth. The### BEGIN SOLUTION/### END SOLUTIONmarkers are intentional nbgrader annotations.When a student runs
tito src export, these blocks get stripped out to generate the student-facing notebook with empty stubs. The GitHub source view always shows the full implementation by design.To get the blank student version locally:
This generates the notebook in
modules/01_tensor/with only the TODO stubs visible. Not a bug, working as intended.@bdub-1 commented on GitHub (May 5, 2026):
Thanks for the response. I used the quick start guide to being my module. I also don't see any mention of the this in the docs unless I missed it during the setup instructions. When I run
tito --helpthere is noexportcommand that shows up.edit: I am following the student-learner role so I was confused why I am seeing the answers as my default. The TITO CLI module workflow page doesn't mention this either. I cannot get the blank version to generate so I am unsure if there actually is a bug. I saw a closed issue earlier of the same bug so I am wondering if the issue lingers
@Shashank-Tripathi-07 commented on GitHub (May 6, 2026):
This is a confirmed bug. When a student runs
tito module start 01, the CLI was generating the notebook by passing the rawsrc/01_tensor/01_tensor.pyfile directly to jupytext with no stripping step. Every### BEGIN SOLUTION/### END SOLUTIONblock was carried over intact into the student notebook, so all answers were already filled in from the start.The issue was in
_create_module_from_srcinsidetinytorch/tito/commands/module/workflow.py. The function's own docstring acknowledged this behavior:The fix strips those solution blocks in memory before conversion. The source files in
src/are never touched. A temp directory is used to hold the stripped version, jupytext converts it to a notebook, and the temp dir is cleaned up automatically. Students now receive blank stubs with TODO comments, as intended.PR: #1686
@Shashank-Tripathi-07 commented on GitHub (May 6, 2026):
Thanks for your attention to this bug @bdub-1 . I've made a PR and corrected this to make sure no future versions suffer from any problem like these !
We highly needed learner feedback for the project so that we can correct it further and better ! Thanks
@bdub-1 commented on GitHub (May 6, 2026):
Happy to help!