Skip to content

Batch

Definition
AI-generated

In neural-network training, a batch (often a minibatch) is the subset of examples whose losses are averaged to form one gradient estimate for an optimizer step; batch size trades off noise in updates against memory use and wall-clock time.

Topics
LLM and Agents Machine learning concepts
Synonyms
Minibatch Mini-batch
Related
Batch Effect Genome-Wide Association Study (GWAS) Gradient Descent

Why it matters in GWAS

Large single-cell or sequence trainers choose batch sizes constrained by GPU RAM; batch-driven batch normalization statistics can couple to biological batch effects if plates or donors are confounded with labels.

Example usage

"We capped the per-step batch at 512 cells after gradient accumulation so effective batch size stayed 4096 without OOM."

Batch Effect Genome-Wide Association Study (GWAS) Gradient Descent Learning Rate Loss Function Optimizer Training Epoch
Related-term network

Click any node to open that term page.

Node order. The current term is first. Next come every page linked from Related terms, in the same top-to-bottom order as that list. Then new terms appear in two-hop layers: for each of those direct neighbors (in that same order), their related terms are appended if not already listed. A third pass does the same for two-hop nodes. The walk stops when a fixed maximum number of nodes is reached.

References

  • Goodfellow I, Bengio Y, Courville A. (2016). Deep Learning. MIT Press.
← Base Quality Score Recalibration (BQSR) Batch Effect →

Last updated (UTC · Git history)