Causal depthwise conv1d in CUDA with a PyTorch interface

Features:

• Support fp32, fp16, bf16.
• Kernel size 2, 3, 4.

from causal_conv1d import causal_conv1d_fn

def causal_conv1d_fn(x, weight, bias=None, activation=None):
 """
 x: (batch, dim, seqlen)
 weight: (dim, width)
 bias: (dim,)
 activation: either None or "silu" or "swish"

 out: (batch, dim, seqlen)
 """

Equivalent to:

import torch.nn.functional as F
F.conv1d(x, weight.unsqueeze(1), bias, padding=width - 1, groups=dim)[..., :seqlen]

If you are on ROCm 6.0, run the following steps to avoid errors during compilation. This is not required for ROCm 6.1 onwards.

1. Locate your ROCm installation directory. This is typically found at /opt/rocm/, but may vary depending on your installation.

2. Apply the Patch. Run with sudo in case you encounter permission issues.

    patch /opt/rocm/include/hip/amd_detail/amd_hip_bf16.h < rocm_patch/rocm6_0.patch