title	TorchCode
emoji	🔥
colorFrom	red
colorTo	yellow
sdk	docker
app_port	7860
pinned	false

🔥 TorchCode

Crack the PyTorch interview.

Practice implementing operators and architectures from scratch — the exact skills top ML teams test for.

An interactive coding platform, but for tensors. Self-hosted. Jupyter-based. Instant feedback.

---

🎯 Why TorchCode?

Top companies (Meta, Google DeepMind, OpenAI, etc.) expect ML engineers to implement core operations from memory on a whiteboard. Reading papers isn't enough — you need to write softmax, LayerNorm, MultiHeadAttention, and full Transformer blocks code.

TorchCode gives you a structured practice environment with:

	Feature
🧩	41 curated problems	The most frequently asked PyTorch interview topics
⚖️	Automated judge	Correctness checks, gradient verification, and timing
🎨	Instant feedback	Colored pass/fail per test case, just like competitive programming
💡	Hints when stuck	Nudges without full spoilers
📖	Reference solutions	Study optimal implementations after your attempt
📊	Progress tracking	What you've solved, best times, and attempt counts
🔄	One-click reset	Toolbar button to reset any notebook back to its blank template — practice the same problem as many times as you want
	Open in Colab	Every notebook has an "Open in Colab" badge + toolbar button — run problems in Google Colab with zero setup

No cloud. No signup. No GPU needed. Just make run — or try it instantly on Hugging Face.

---

🚀 Quick Start

Option 0 — Try it online (zero install)

Launch on Hugging Face Spaces — opens a full JupyterLab environment in your browser. Nothing to install.

Or open any problem directly in Google Colab — every notebook has an badge.

Option 0b — Use the judge in Colab (pip)

In Google Colab, install the judge from PyPI so you can run check(...) without cloning the repo:

!pip install torch-judge

Then in a notebook cell:

from torch_judge import check, status, hint, reset_progress
status()           # list all problems and your progress
check("relu")      # run tests for the "relu" task
hint("relu")       # show a hint

Option 1 — Pull the pre-built image (fastest)

docker run -p 8888:8888 -e PORT=8888 ghcr.io/duoan/torchcode:latest

If the registry image is unavailable for your platform, use Option 2 instead. This is the common path on Apple Silicon / arm64.

Option 2 — Build locally

make run

make run will try the prebuilt image first and automatically fall back to a local build when needed.

Open http://localhost:8888 — that's it. Works with both Docker and Podman (auto-detected).

Option 3 — Standalone Web UI (Next.js + FastAPI)

With Python 3.10+ and Node.js 20.9+ (npm) installed, run this from the repository root to launch the coding workspace and daily practice dashboard together:

./start.sh

The script creates a Python virtual environment if needed, installs API and frontend dependencies, and starts:

• http://localhost:4000 — your daily practice dashboard.
• http://localhost:3000 — the coding workspace with an integrated IDE, debugger, and test feedback.
• http://localhost:8000 — the FastAPI backend shared by both UIs.

On the localhost:4000 dashboard:

• Today’s practice: choose a daily goal (3 problems by default), mixing new lessons in Suggested Study Plan order with timely reviews.
• Review lab & Your journey: revisit tricky problems, explore four weekly paths, and view submission history grouped by problem with test results.
• Practice It! & Solution: open a problem on localhost:3000 or view its solution code and explanations, including interactive examples and a fullscreen option.

To stop all three services, run this from the repository root in another terminal:

./stop.sh

You can also press Ctrl+C in the launch terminal. Ports 3000, 4000, and 8000 must be available. If dependencies are already installed, use SKIP_INSTALL=1 ./start.sh to skip installation.

Practice features and saved history

Choose a local learner profile, adjust the daily problem count, and click Practice It! to open the selected problem in the coding workspace. Submitting code saves the attempt and test results; the dashboard uses these results to prioritize unfinished problems and schedule review of past problems. Each profile keeps its own goal and history, stored locally in SQLite under data/ and retained across restarts.

Your Journey turns the Suggested Study Plan below into four illustrated weekly paths, with all 41 problems in study order. Select a problem to see your progress and open its practice workspace or solution. Once a problem has a correct submission, it stays Solved and counts toward Journey completion, even if you choose Needs more practice or make mistakes in later attempts. The practice flag remains a separate reminder for future recommendations. Practice history groups repeated submissions under each problem; expand a row to see individual attempts and their test results.

In Your daily mission, a problem stays Done today once you pass its tests that day, even if you choose Needs more practice or try again unsuccessfully. The flag helps schedule future practice without taking away earned completion. On a later day, a recommended review needs a fresh successful submission to count toward that day's mission.

Reviews on Today’s practice and Review lab are ranked by a priority score from 0–100: 40% your review flag + 35% recent submission mistakes + 25% estimated forgetting. The mistake signal uses the last five submissions, weights newer attempts more heavily, and includes the fraction of tests missed. Clean practice lowers the score; confident revisits increase the time before the next review. New problems continue to follow the Suggested Study Plan, and changing a practice flag updates recommendations without removing today's earned completion.

The forgetting estimate uses recall = 2^(-days_since_success / half_life), following the half-life curve described in Duolingo’s spaced-repetition research. Memory strength grows with spaced clean passes and decreases with recent mistakes. Failed submissions and flag clicks do not reset the last successful-practice timestamp. These weights and memory estimates are transparent heuristics, not a trained or calibrated measure of an individual’s recall; older imported completion markers without dated submissions do not receive an invented practice timestamp.

The Solution button beside each practice button opens a guide with Solution Code and Explanation tabs. Use Fullscreen in the guide's header to fill the browser window; Exit fullscreen or Esc returns to the smaller view without losing your place. Code comes from the reference notebook; explanations for all 41 problems are pre-generated in explanations/. Each guide walks through the actual code, shows intermediate values in worked examples and diagrams, and explains the reference's edge cases. Selected guides also include interactive examples for dropout, embedding, attention, masks, normalization, RoPE, sampling, and PPO. The code tab supports copying the implementation, including its imports, back into the editor.

In the coding workspace, Submission history shows the current learner's past attempts for the selected problem, including submission times and passed/total test counts. Click an attempt to load its exact code into the editor; Undo load restores the previous draft. New submissions save source code locally alongside test results. Older attempts retain their results but show Code unavailable because their source code was not recorded.

Open the coding workspace once to import any existing browser-saved solved markers. These markers preserve completion status; detailed history starts with recorded submissions. Profiles are local practice identities, with no signup or account authentication.

---

📋 Problem Set

Frequency: 🔥 = very likely in interviews, ⭐ = commonly asked, 💡 = emerging / differentiator

🧱 Fundamentals — "Implement X from scratch"

The bread and butter of ML coding interviews. You'll be asked to write these without torch.nn.

#	Problem	What You'll Implement	Difficulty	Freq	Key Concepts
1	ReLU	relu(x)		🔥	Activation functions, element-wise ops
2	Softmax	my_softmax(x, dim)		🔥	Numerical stability, exp/log tricks
16	Cross-Entropy Loss	cross_entropy_loss(logits, targets)		🔥	Log-softmax, logsumexp trick
17	Dropout	MyDropout (nn.Module)		🔥	Train/eval mode, inverted scaling
18	Embedding	MyEmbedding (nn.Module)		🔥	Lookup table, weight[indices]
19	GELU	my_gelu(x)		⭐	Gaussian error linear unit, torch.erf
20	Kaiming Init	kaiming_init(weight)		⭐	std = sqrt(2/fan_in), variance scaling
21	Gradient Clipping	clip_grad_norm(params, max_norm)		⭐	Norm-based clipping, direction preservation
31	Gradient Accumulation	accumulated_step(model, opt, ...)		💡	Micro-batching, loss scaling
40	Linear Regression	LinearRegression (3 methods)		🔥	Normal equation, GD from scratch, nn.Linear
3	Linear Layer	SimpleLinear (nn.Module)		🔥	y = xW^T + b, Kaiming init, nn.Parameter
4	LayerNorm	my_layer_norm(x, γ, β)		🔥	Normalization, running stats, affine transform
7	BatchNorm	my_batch_norm(x, γ, β)		⭐	Batch vs layer statistics, train/eval behavior
8	RMSNorm	rms_norm(x, weight)		⭐	LLaMA-style norm, simpler than LayerNorm
15	SwiGLU MLP	SwiGLUMLP (nn.Module)		⭐	Gated FFN, SiLU(gate) * up, LLaMA/Mistral-style
22	Conv2d	my_conv2d(x, weight, ...)		🔥	Convolution, unfold, stride/padding

🧠 Attention Mechanisms — The heart of modern ML interviews

If you're interviewing for any role touching LLMs or Transformers, expect at least one of these.

#	Problem	What You'll Implement	Difficulty	Freq	Key Concepts
23	Cross-Attention	MultiHeadCrossAttention (nn.Module)		⭐	Encoder-decoder, Q from decoder, K/V from encoder
5	Scaled Dot-Product Attention	scaled_dot_product_attention(Q, K, V)		🔥	softmax(QK^T/√d_k)V, the foundation of everything
6	Multi-Head Attention	MultiHeadAttention (nn.Module)		🔥	Parallel heads, split/concat, projection matrices
9	Causal Self-Attention	causal_attention(Q, K, V)		🔥	Autoregressive masking with -inf, GPT-style
10	Grouped Query Attention	GroupQueryAttention (nn.Module)		⭐	GQA (LLaMA 2), KV sharing across heads
11	Sliding Window Attention	sliding_window_attention(Q, K, V, w)		⭐	Mistral-style local attention, O(n·w) complexity
12	Linear Attention	linear_attention(Q, K, V)		💡	Kernel trick, φ(Q)(φ(K)^TV), O(n·d²)
14	KV Cache Attention	KVCacheAttention (nn.Module)		🔥	Incremental decoding, cache K/V, prefill vs decode
24	RoPE	apply_rope(q, k)		🔥	Rotary position embedding, relative position via rotation
25	Flash Attention	flash_attention(Q, K, V, block_size)		💡	Tiled attention, online softmax, memory-efficient

🏗️ Architecture & Adaptation — Put it all together

#	Problem	What You'll Implement	Difficulty	Freq	Key Concepts
26	LoRA	LoRALinear (nn.Module)		⭐	Low-rank adaptation, frozen base + BA update
27	ViT Patch Embedding	PatchEmbedding (nn.Module)		💡	Image → patches → linear projection
13	GPT-2 Block	GPT2Block (nn.Module)		⭐	Pre-norm, causal MHA + MLP (4x, GELU), residual connections
28	Mixture of Experts	MixtureOfExperts (nn.Module)		⭐	Mixtral-style, top-k routing, expert MLPs

⚙️ Training & Optimization

#	Problem	What You'll Implement	Difficulty	Freq	Key Concepts
29	Adam Optimizer	MyAdam		⭐	Momentum + RMSProp, bias correction
30	Cosine LR Scheduler	cosine_lr_schedule(step, ...)		⭐	Linear warmup + cosine annealing

🎯 Inference & Decoding

#	Problem	What You'll Implement	Difficulty	Freq	Key Concepts
32	Top-k / Top-p Sampling	sample_top_k_top_p(logits, ...)		🔥	Nucleus sampling, temperature scaling
33	Beam Search	beam_search(log_prob_fn, ...)		🔥	Hypothesis expansion, pruning, eos handling
34	Speculative Decoding	speculative_decode(target, draft, ...)		💡	Accept/reject, draft model acceleration

🔬 Advanced — Differentiators

#	Problem	What You'll Implement	Difficulty	Freq	Key Concepts
35	BPE Tokenizer	SimpleBPE		💡	Byte-pair encoding, merge rules, subword splits
36	INT8 Quantization	Int8Linear (nn.Module)		💡	Per-channel quantize, scale/zero-point, buffer vs param
37	DPO Loss	dpo_loss(chosen, rejected, ...)		💡	Direct preference optimization, alignment training
38	GRPO Loss	grpo_loss(logps, rewards, group_ids, eps)		💡	Group relative policy optimization, RLAIF, within-group normalized advantages
39	PPO Loss	ppo_loss(new_logps, old_logps, advantages, clip_ratio)		💡	PPO clipped surrogate loss, policy gradient, trust region
41	OPD Loss	opd_loss(student_logits, teacher_logits, ...)		💡	On-policy distillation, reverse KL, multi-teacher alignment

---

⚙️ How It Works

Each problem has two notebooks:

File	Purpose
01_relu.ipynb	✏️ Blank template — write your code here
01_relu_solution.ipynb	📖 Reference solution — check when stuck

Workflow

1. Open a blank notebook           →  Read the problem description
2. Implement your solution         →  Use only basic PyTorch ops
3. Debug freely                    →  print(x.shape), check gradients, etc.
4. Run the judge cell              →  check("relu")
5. See instant colored feedback    →  ✅ pass / ❌ fail per test case
6. Stuck? Get a nudge              →  hint("relu")
7. Review the reference solution   →  01_relu_solution.ipynb
8. Click 🔄 Reset in the toolbar  →  Blank slate — practice again!

In-Notebook API

from torch_judge import check, hint, status

check("relu")               # Judge your implementation
hint("causal_attention")    # Get a hint without full spoiler
status()                    # Progress dashboard — solved / attempted / todo

---

📅 Suggested Study Plan

All 41 problems: ~12–16 hours spread across 3–4 weeks. Perfect for interview prep on a deadline.

Week	Focus	Problems	Time
1	🧱 Foundations	ReLU → Softmax → CE Loss → Dropout → Embedding → GELU → Linear → Linear Regression → LayerNorm → BatchNorm → RMSNorm → SwiGLU MLP → Conv2d	2–3 hrs
2	🧠 Attention Deep Dive	SDPA → MHA → Cross-Attn → Causal → GQA → KV Cache → Sliding Window → RoPE → Linear Attn → Flash Attn	3–4 hrs
3	🏗️ Architecture + Training	GPT-2 Block → LoRA → MoE → ViT Patch → Adam → Cosine LR → Grad Clip → Grad Accumulation → Kaiming Init	3–4 hrs
4	🎯 Inference + Advanced	Top-k/p Sampling → Beam Search → Speculative Decoding → BPE → INT8 Quant → DPO Loss → GRPO Loss → PPO Loss → OPD Loss + speed run	3–4 hrs

---

🏛️ Architecture

┌──────────────────────────────────────────┐
│           Docker / Podman Container      │
│                                          │
│  JupyterLab (:8888)                      │
│    ├── templates/  (reset on each run)   │
│    ├── solutions/  (reference impl)      │
│    ├── torch_judge/ (auto-grading)       │
│    ├── torchcode-labext (JLab plugin)    │
│    │     🔄 Reset — restore template     │
│    │     🔗 Colab — open in Colab        │
│    └── PyTorch (CPU), NumPy              │
│                                          │
│  Judge checks:                           │
│    ✓ Output correctness (allclose)       │
│    ✓ Gradient flow (autograd)            │
│    ✓ Shape consistency                   │
│    ✓ Edge cases & numerical stability    │
└──────────────────────────────────────────┘

Single container. Single port. No database. No frontend framework. No GPU.

🛠️ Commands

make run    # Build & start (http://localhost:8888)
make stop   # Stop the container
make clean  # Stop + remove volumes + reset all progress

🧩 Adding Your Own Problems

TorchCode uses auto-discovery — just drop a new file in torch_judge/tasks/:

TASK = {
    "id": "my_task",
    "title": "My Custom Problem",
    "difficulty": "medium",
    "function_name": "my_function",
    "hint": "Think about broadcasting...",
    "tests": [ ... ],
}

No registration needed. The judge picks it up automatically.

---

📦 Publishing torch-judge to PyPI (maintainers)

The judge is published as a separate package so Colab/users can pip install torch-judge without cloning the repo.

Automatic (GitHub Action)

Pushing to master after changing the package version triggers .github/workflows/pypi-publish.yml, which builds and uploads to PyPI. No git tag is required.

1. Bump version in torch_judge/_version.py (e.g. __version__ = "0.1.1").
2. Configure PyPI Trusted Publisher (one-time):
   • PyPI → Your project torch-judge → Publishing → Add a new pending publisher
   • Owner: duoan, Repository: TorchCode, Workflow: pypi-publish.yml, Environment: (leave empty)
   • Run the workflow once (push a version bump to master or Actions → Publish torch-judge to PyPI → Run workflow); PyPI will then link the publisher.
3. Release: commit the version bump and git push origin master.

Alternatively, use an API token: add repository secret PYPI_API_TOKEN (value = pypi-... from PyPI) and set TWINE_USERNAME=__token__ and TWINE_PASSWORD from that secret in the workflow if you prefer not to use Trusted Publishing.

Manual

pip install build twine
python -m build
twine upload dist/*

Version is in torch_judge/_version.py; bump it before each release.

---

❓ FAQ

Do I need a GPU?

No. Everything runs on CPU. The problems test correctness and understanding, not throughput.

Can I keep my solutions between runs?

Blank templates reset on every make run so you practice from scratch. Save your work under a different filename if you want to keep it. You can also click the 🔄 Reset button in the notebook toolbar at any time to restore the blank template without restarting.

Can I use Google Colab instead?

Yes! Every notebook has an Open in Colab badge at the top. Click it to open the problem directly in Google Colab — no Docker or local setup needed. You can also use the Colab toolbar button inside JupyterLab.

How are solutions graded?

The judge runs your function against multiple test cases using torch.allclose for numerical correctness, verifies gradients flow properly via autograd, and checks edge cases specific to each operation.

Who is this for?

Anyone preparing for ML/AI engineering interviews at top tech companies, or anyone who wants to deeply understand how PyTorch operations work under the hood.

---

🤝 Contributors

Thanks to everyone who has contributed to TorchCode.

duoan	Ando233	abhijitmjj	laitifranz	hrlics	HareshKarnan
ThierryHJ	Zuozhuo	reidemeister94

Auto-generated from the GitHub contributors graph with avatars and GitHub usernames.

---

Built for engineers who want to deeply understand what they build.

If this helped your interview prep, consider giving it a ⭐

---

☕ Buy Me a Coffee

Scan to support