Top k sampling from probs v163840
0 eligible runs
sampling
Top-k sampling from probabilities with vocab_size=163840. Keeps only the k highest probability tokens, renormalizes, then samples from the filtered distribution. Captured from Kimi K2.5 (moonshotai/Kimi-K2.5, text_config.vocab_size=163840).
Current records
No published measurement for the selected workload.
Implementations
Implementation
Runtime
Best latency
Evidence
Availability
Semantics
Inputs and outputs
probsfp32 [batch_size, vocab_size]
top_kint32 [batch_size]
samplesint64 [batch_size]
Axes and behavior
batch_sizevariable
vocab_sizeconstant = 163840
determinismunspecified
constraintsNo mutation or aliasing
Identity
No source imports for this operation yet.How records are decidedJSON