For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Mojo function

ssd_combined_gpu

ssd_combined_gpu[kernel_dtype: DType, DSTATE: Int, output_layout: Layout, x_layout: Layout, out_z_layout: Layout, residual_layout: Layout, u_layout: Layout, delta_layout: Layout, A_layout: Layout, B_layout: Layout, C_layout: Layout, D_layout: Layout, z_layout: Layout, delta_bias_layout: Layout, gamma_layout: Layout](total_batch_dim: Int, batch: Int, dim: Int, seqlen: Int, group_size: Int, delta_softplus: Int8, output: LayoutTensor[kernel_dtype, output_layout, MutAnyOrigin], x: LayoutTensor[kernel_dtype, x_layout, MutAnyOrigin], out_z: LayoutTensor[kernel_dtype, out_z_layout, MutAnyOrigin], residual: LayoutTensor[kernel_dtype, residual_layout, MutAnyOrigin], u: LayoutTensor[kernel_dtype, u_layout, MutAnyOrigin], delta: LayoutTensor[kernel_dtype, delta_layout, MutAnyOrigin], A: LayoutTensor[kernel_dtype, A_layout, MutAnyOrigin], B: LayoutTensor[kernel_dtype, B_layout, MutAnyOrigin], C: LayoutTensor[kernel_dtype, C_layout, MutAnyOrigin], D: LayoutTensor[kernel_dtype, D_layout, MutAnyOrigin], z: LayoutTensor[kernel_dtype, z_layout, MutAnyOrigin], delta_bias: LayoutTensor[kernel_dtype, delta_bias_layout, MutAnyOrigin], gamma: LayoutTensor[kernel_dtype, gamma_layout, MutAnyOrigin], epsilon: Scalar[kernel_dtype], weight_offset: Scalar[kernel_dtype])

GPU kernel for SSD combined operation.

Combines selective scan with normalization and residual connection. Performs: norm(residual + selective_scan(input))

Was this page helpful?

Thank you! We'll create more content like this.

Thank you for helping us improve!

😔 What went wrong?

Some code doesn’t work

It includes inaccurate information

It's missing information I need

It was difficult to understand

Other

Submit