Sign inSign in
ConceptAI & ML

Gradient accumulation

Accumulating gradients via addition in backpropagation ensures that variables used multiple times correctly sum their derivative contributions.
Watch it explained

Know first

Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.

Backpropagation2Optimizer step2Backward pass1Mini-batch training1Multivariate chain rule1

Opens up

Concepts that name this one as a prerequisite.

Lecturers

1
  • Andrej Karpathy

    Undergraduate

    • 2:36:41Let's Reproduce GPT-2 (124M): Distributed PyTorch, FlashAttention & FineWeb

Connections

6

The lectures that define it or assume it, and the concepts next to it.

1 lecturer. Definitions are written by AI from each lecture’s transcript. A time with a play mark starts the lecture at the moment the lecturer says it.