Rustlab LLM

README.md
00-the-llm-as-a-system.md
01-tokens-and-encoding.md
02-probability-and-softmax.md
03-cross-entropy-loss.md
04-embeddings-and-similarity.md
05-bigram-language-model.md
06-linear-layers-and-gradient-descent.md
07-context-and-naive-averaging.md
08-scaled-dot-product-attention.md
09-multi-head-attention.md
10-positional-encoding.md
11-feed-forward-block.md
12-layer-norm-and-residuals.md
13-transformer-block.md
14-full-gpt-architecture.md
15-backpropagation.md
16-adamw-optimizer.md
17-learning-rate-scheduling.md
18-training-loop.md
19-byte-pair-encoding.md
20-perplexity-and-evaluation.md
21-sampling-and-generation.md
22-full-backprop-through-the-block.md
23-putting-it-all-together.md
24-modern-architectural-variants.md
25-fine-tuning-sft-and-dpo.md
26-quantization-and-fixed-point-inference.md

Rustlab LLM

28 notebooks

rustlab_llm — Lesson Site
Lesson 00: The LLM as a System
Lesson 01: Tokens & Text Encoding
Lesson 02: Probability & Softmax
Lesson 03: Cross-Entropy Loss
Lesson 04: Embeddings & Similarity
Lesson 05: The Bigram Language Model
Lesson 06: Linear Layers & Gradient Descent
Lesson 07: Context and Naive Averaging
Lesson 08: Scaled Dot-Product Attention
Lesson 09: Multi-Head Attention
Lesson 10: Positional Encoding
Lesson 11: Feed-Forward Block (MLP)
Lesson 12: LayerNorm & Residual Connections
Lesson 13: The Transformer Block
Lesson 14: Full GPT Architecture
Lesson 15: Backpropagation
Lesson 16: The AdamW Optimizer
Lesson 17: Learning Rate Scheduling
Lesson 18: The Training Loop
Lesson 19: Byte-Pair Encoding (BPE)
Lesson 20: Perplexity and Evaluation
Lesson 21: Sampling and Generation
Lesson 22: Full Backprop Through the Block
Lesson 23: Putting It All Together
Lesson 24: Modern Architectural Variants
Lesson 25: Fine-Tuning — SFT and DPO
Lesson 26: Quantization and Fixed-Point Inference

Generated by rustlab-notebook