[ 仅 API 调用方式不一致 ]torch.nn.MultiheadAttention

torch.nn.MultiheadAttention

torch.nn.MultiheadAttention(embed_dim, num_heads, dropout=0.0, bias=True, add_bias_kv=False, add_zero_attn=False, kdim=None, vdim=None, batch_first=False, device=None, dtype=None)

paddle.compat.nn.MultiheadAttention

paddle.compat.nn.MultiheadAttention(embed_dim, num_heads, dropout=0.0, bias=True, add_bias_kv=False, add_zero_attn=False, kdim=None, vdim=None, batch_first=False, device=None, dtype=None)

两者功能和用法一致,但 API 路径不一致,只需修改 torch 前缀为 paddle.compat,具体如下:

转写示例

# PyTorch 写法
layer = torch.nn.MultiheadAttention(128, 8)

# Paddle 写法
layer = paddle.compat.nn.MultiheadAttention(128, 8)