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).
Python function
short_conv_ring_fwd
short_conv_ring_fwd()
max.nn.state_space.short_conv_ring_fwd(x, weight, ring, input_row_offsets, positions, conv_row)
Returns x + conv(x) over a ragged batch; reads ring, writes
nothing.
-
Parameters:
-
- x (TensorValue) –
[total_seq_len, channels]input. - weight (TensorValue) –
[channels, kernel_size]taps; the last multiplies the current token. - ring (BufferValue) –
[slots, ring_len, channels]conv state. - input_row_offsets (TensorValue) –
[batch + 1]uint32. - positions (TensorValue) –
[total_seq_len]uint32 position per token. - conv_row (TensorValue) –
[batch]uint32 ring slot per sequence.
- x (TensorValue) –
-
Returns:
-
Same shape and dtype as
x. -
Return type: