mirror of https://github.com/MLSysBook/TinyTorch.git synced 2026-03-11 21:43:34 -05:00

Files

Vijay Janapa Reddi 97e0563614 Add community and benchmark features with baseline validation

- Implement tito benchmark baseline and capstone commands
- Add SPEC-style normalization for baseline benchmarks
- Implement tito community join, update, leave, stats, profile commands
- Use project-local storage (.tinytorch/) for user data
- Add privacy-by-design with explicit consent prompts
- Update site documentation for community and benchmark features
- Add Marimo integration for online notebooks
- Clean up redundant milestone setup exploration docs
- Finalize baseline design: fast setup validation (~1 second) with normalized results

2025-11-20 00:17:21 -05:00

2.8 KiB

Raw Blame History

TinyTorch Module Order Verification

✅ Correct Module Order (modules/ directory)

01_tensor          - Foundation: N-dimensional arrays
02_activations     - Non-linearity functions (ReLU, Sigmoid, Softmax)
03_layers         - Neural network layers (Linear, Module base)
04_losses         - Loss functions (MSE, CrossEntropy)
05_autograd       - Automatic differentiation
06_optimizers     - Optimization algorithms (SGD, Adam)
07_training       - Training loops
08_dataloader     - Data batching and pipelines
09_spatial        - Convolutional operations
10_tokenization   - Text tokenization
11_embeddings     - Word embeddings
12_attention      - Attention mechanisms
13_transformers   - Transformer architecture
14_profiling      - Performance profiling
15_quantization   - Model quantization
16_compression    - Model compression
17_memoization    - KV caching
18_acceleration   - Hardware acceleration
19_benchmarking   - Performance benchmarking
20_capstone       - Torch Olympics competition

⚠️ Issue Found: Assignments Directory Mismatch

Current assignments/source/ structure:

01_setup    ❌ OUTDATED - Module 01 is now "tensor", not "setup"
02_tensor   ✅ Correct

Problem: The assignments/source/01_setup/ directory contains an old notebook from when Module 01 was "Setup". Module 01 is now "Tensor" (modules/01_tensor/).

Impact on Binder/Colab

No impact - Binder setup doesn't depend on assignment notebooks. The binder/ configuration:

Installs TinyTorch package (pip install -e .)
Provides JupyterLab environment
Students can access any notebooks in the repository

However, for consistency and to avoid confusion:

Old 01_setup assignment should be removed or renamed
Documentation references should point to 01_tensor (already fixed)

Module Tiers (from site/_toc.yml)

🏗️ Foundation Tier (01-07)

01 Tensor
02 Activations
03 Layers
04 Losses
05 Autograd
06 Optimizers
07 Training

🏛️ Architecture Tier (08-13)

08 DataLoader
09 Spatial (Convolutions)
10 Tokenization
11 Embeddings
12 Attention
13 Transformers

⏱️ Optimization Tier (14-19)

14 Profiling
15 Quantization
16 Compression
17 Memoization
18 Acceleration
19 Benchmarking

🏅 Capstone (20)

20 Capstone (Torch Olympics)

Verification Status

✅ Modules directory: Correct order (01-20) ✅ Documentation: References updated to 01_tensor ✅ Binder setup: Not affected by assignment structure ⚠️ Assignments: Contains outdated 01_setup (should be removed)

Recommendations

Remove old assignment: Delete assignments/source/01_setup/ and assignments/release/01_setup/
Verify nbgrader: Ensure nbgrader commands reference correct module numbers
Update any remaining references: Search for 01_setup and update to 01_tensor

2.8 KiB Raw Blame History