factjuice
Technology  /  transformers

Facts about Transformers (in AI)

7 facts squeezed so far
  1. 07

    In 2023, researchers at Google introduced FlashAttention, reducing transformer attention computation from O(n²) to O(n log n) complexity by computing attention in GPU memory blocks rather than full matrices.

    Transformers (in AI)May 14efficiencyoptimizationalgorithm
  2. 06

    Multi-head attention in transformers splits the embedding dimension into 8 to 12 parallel attention heads, allowing the model to simultaneously attend to information from different representation subspaces.

    Transformers (in AI)May 14architectureneuralnetworksmechanism
  3. 05

    Positional encoding in transformers uses sinusoidal functions with wavelengths ranging from 2π to 10000·2π to allow models to extrapolate to sequence lengths longer than those seen during training.

    Transformers (in AI)May 14architecturemathematicstraining
  4. 04

    Transformer models require O(n²) memory during training due to computing full attention matrices, making it prohibitively expensive to train on sequences longer than 2048 tokens without techniques like gradient checkpointing or sparse attention patterns.

    Transformers (in AI)May 14architecturecomputationallimitation
  5. 03

    During inference, transformer models use key-value caching to reduce computational complexity from O(n²) to O(n) per token, enabling practical deployment of billion-parameter models in production systems.

    Transformers (in AI)May 14optimizationcomputationefficiency
  6. 02

    GPT-3, released by OpenAI in 2020, contains 175 billion parameters and demonstrates that transformer models can perform novel tasks without task-specific fine-tuning through in-context learning.

    Transformers (in AI)May 14scalinglanguageneuralnetworks
  7. 01

    The 2017 Attention Is All You Need paper introduced transformers with a 65,000-parameter base model that outperformed recurrent neural networks on machine translation tasks.

    Transformers (in AI)May 14architecturebreakthroughmachine-learning