Problems

Difficulty:

Category:

Tag:

Title	Difficulty	Category
MLM Eval — Masked Accuracy	medium	end_to_end
MLM Forward Pass	medium	end_to_end
Adapter Layer	medium	end_to_end
Training Loop	easy	end_to_end
Logistic Regression From Scratch	easy	end_to_end
Data Augmentation	medium	end_to_end
Transformer Encoder Block	hard	end_to_end
Multi-Class Classifier	easy	end_to_end
Gradient Clipping	easy	end_to_end
Transformer Decoder Block	hard	end_to_end
BPE Encode Text	hard	end_to_end
Feature Normalization Pipeline	easy	end_to_end
Eval Loop with Metrics	medium	end_to_end
Gradient Accumulation	medium	end_to_end
Mini-Batch Training	medium	end_to_end
Beam Search	hard	end_to_end
Model Checkpointing	easy	end_to_end
MLM Masking Strategy	medium	end_to_end
Word Embedding Model	medium	end_to_end
Polynomial Regression	medium	end_to_end
Autoencoder	medium	end_to_end
Linear Regression	easy	end_to_end
Tokenize and Pad Batch	medium	end_to_end
Weight Initialization	easy	end_to_end
ViT with Mixup Augmentation	medium	end_to_end
Simple RNN Cell	medium	end_to_end
Binary Classifier	easy	end_to_end
Causal Self-Attention Block	medium	end_to_end
Distributed Training Step End-to-End	hard	end_to_end
DDPM Sampler Loop	hard	end_to_end
DDPM Training Step End-to-End	medium	end_to_end
Simple GAN Generator	hard	end_to_end
Learning Rate Scheduler	medium	end_to_end
Multi-Head Attention Block	hard	end_to_end
Train Binary Classifier End-to-End	medium	end_to_end
Nucleus Sampling Generation Loop	hard	end_to_end
Two-Layer MLP	easy	end_to_end
GRU Cell	medium	end_to_end
Train Image Classifier End-to-End	medium	end_to_end
Train LoRA Adapter End-to-End	medium	end_to_end
LSTM Cell	medium	end_to_end
Train Multiclass Classifier with Early Stopping	medium	end_to_end
Train RNN Language Model	hard	end_to_end
Train VAE End-to-End	medium	end_to_end
Train with Adam End-to-End	medium	end_to_end
Skip Connection Block	medium	end_to_end
Transformer Block with RoPE	hard	end_to_end
Causal LM Forward Pass	medium	end_to_end
Sequence Classifier	medium	end_to_end
Self-Attention Layer	hard	end_to_end
LLaMA-Style Transformer Block	hard	end_to_end
Causal LM with KV Cache Generation	hard	end_to_end
Cross-Attention Block	medium	end_to_end
Encoder-Decoder Transformer Forward Pass	hard	end_to_end
Simple CNN	medium	end_to_end
Encoder-Decoder Greedy Decode	medium	end_to_end
Encoder-Decoder Beam Search	hard	end_to_end
MLM Forward with Tied Output Head	medium	end_to_end
Train ViT Classifier End-to-End	hard	end_to_end
ViT Classification Forward	hard	end_to_end
ViT Encoder Block	medium	end_to_end
Patch Embedding with CLS Token	medium	end_to_end
Train Causal LM Pretraining Step	hard	end_to_end
Train Encoder-Decoder Seq2Seq Step	hard	end_to_end
Train MLM Pretraining Step	hard	end_to_end
Train Tiny GPT End-to-End	hard	end_to_end
Pure Function vs Impure	easy	end_to_end
Jit with static_argnames	medium	end_to_end
Tracer Leak Detection	medium	end_to_end
Tracing: Shape vs Value	medium	end_to_end