For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).
Mojo struct
CausalPaddingMask
struct CausalPaddingMask[LayoutType: TensorLayout, origin_: ImmOrigin]
Causal mask combined with padding: a position (seq_id, head, q, k) is visible only when q >= k (causal) AND k < valid_lengths[seq_id] (padding).
valid_lengths is a tensor of shape [num_seqs] with one uint32 value per
sequence indicating the number of valid (non-padding) tokens.
Fields
- valid_lengths (
ImmTileTensor[.uint32, LayoutType, origin_]):
Implemented traits
AnyType,
Copyable,
Deinitable,
DevicePassable,
ImplicitlyCopyable,
MHAMask,
Movable,
RegisterPassable,
TrivialRegisterPassable
comptime members
apply_log2e_after_mask
comptime apply_log2e_after_mask = False
causal_mask
comptime causal_mask = CausalMask()
check_mask_during_decoding
comptime check_mask_during_decoding = True
device_type
comptime device_type = CausalPaddingMask[LayoutType, origin_]
has_row_bias
comptime has_row_bias = False
Whether add_row_bias adds this mask's additive bias, leaving visibility_mask the rest of mask.
mask_out_of_bound
comptime mask_out_of_bound = is_nvidia_gpu()
mask_safe_out_of_bounds
comptime mask_safe_out_of_bounds = True
MaskType
comptime MaskType[*Ts: Fused & Input & LoadAccess] = CausalPaddingMask[LayoutType, origin_]
Parameters
- *Ts (
Fused&Input&LoadAccess):
supports_decode_token_fold
comptime supports_decode_token_fold = False
Whether AMD decode may fold several query tokens into one tile under this mask, and start its split-K key range at start_column: both status and start_column stay valid when asked about the span of query rows a sequence's decode step carries.
Methods
__init__
def __init__(valid_lengths: ImmTileTensor[.uint32, LayoutType, origin_]) -> Self
get_type_name
name
mask
def mask[dtype: DType, width: SIMDLength, //](self, coord: Coord[Int, Int, Int, Int], score_vec: SIMD[dtype, width]) -> SIMD[dtype, width]
Returns:
status
def status(self, seq_id: UInt32, tile_offset: Coord[Int, Int], tile_size: Coord[Int, Int]) -> TileMaskStatus
Returns:
start_column
def start_column[BM: Int, BN: Int, page_size: Int](self, seq_id: UInt32, row: UInt32) -> UInt32
Returns:
start_column_alignment
total_iters
def total_iters[BM: Int, BN: Int, page_size: Int](self, seq_id: UInt32, row: UInt32, num_cols: UInt32) -> UInt32
Returns:
count_nonfull_sets
last_masked_set_end
def last_masked_set_end[BM: Int, BN: Int, page_size: Int](self, seq_id: UInt32, row: UInt32, num_cols: UInt32) -> UInt32
Returns:
masked_set_ends
def masked_set_ends[BM: Int, BN: Int, page_size: Int](self, seq_id: UInt32, row: UInt32, num_cols: UInt32) -> StaticTuple[UInt32, CausalPaddingMask.count_nonfull_sets(BM, BN)]
Returns:
StaticTuple[UInt32, CausalPaddingMask.count_nonfull_sets(BM, BN)]
nonfull_sets
static def nonfull_sets[BM: Int, BN: Int]() -> StaticTuple[TileMaskStatus, CausalPaddingMask.count_nonfull_sets(BM, BN)]
Returns:
StaticTuple[TileMaskStatus, CausalPaddingMask.count_nonfull_sets(BM, BN)]
mask_strategies
static def mask_strategies[BM: Int, BN: Int]() -> StaticTuple[MaskStrategy, CausalPaddingMask.count_nonfull_sets(BM, BN)]
Returns:
StaticTuple[MaskStrategy, CausalPaddingMask.count_nonfull_sets(BM, BN)]
mask_bits
def mask_bits(self, seq_id: UInt32, score_row: Int32, col_start: Int32, num_keys: Int32) -> UInt32
Returns: