flashinfer.gdn_paged.GDNPagedDecodeSlots

class flashinfer.gdn_paged.GDNPagedDecodeSlots(initial_state_indices: Tensor, output_state_indices: Tensor)

Per-sequence pool rows of a paged decode batch, both [B] int32. See gdn_paged_decode_slots().

__init__()

Methods

__init__()
count(value, /)	Return number of occurrences of value.
index(value[, start, stop])	Return first index of value.

Attributes

initial_state_indices	Page read as the state before the decoded token.
output_state_indices	Last page, written with the state after the decoded token.