大模型中常用的注意力机制GQA详解以及Pytorch代码实现

from torch.nn.functional import scaled_dot_product_attention  # shapes: (batch_size, seq_len, num_heads, head_dim) query = torch.randn(1, 256, 8, 64) key = torch.randn(1, 256, 8, 64) value = torch.randn(1, 256, 8, 64)  output = scaled_dot_product_attention(query, key, value) print(output.shape) # torch.Size([1, 256, 8, 64])

import torch  # shapes: (batch_size, seq_len, num_heads, head_dim) query = torch.randn(1, 256, 8, 64) key = torch.randn(1, 256, 2, 64) value = torch.randn(1, 256, 2, 64)  num_head_groups = query.shape[2] // key.shape[2] print(num_head_groups) # each group is of size 4 since there are 2 kv_heads

from einops import rearrange  query = rearrange(query, "b n h d -> b h n d") key = rearrange(key, "b s h d -> b h s d") value = rearrange(value, "b s h d -> b h s d")

from einops import rearrange query = rearrange(query, "b (h g) n d -> b g h n d", g=num_head_groups) print(query.shape) # torch.Size([1, 4, 2, 256, 64])

from einops import einsum, rearrange # g stands for the number of groups # h stands for the hidden dim # n and s are equal and stands for sequence length scores = einsum(query, key, "b g h n d, b h s d -> b h n s") print(scores.shape) # torch.Size([1, 2, 256, 256])

import torch.nn.functional as F  scale = query.size(-1) ** 0.5 attention = F.softmax(similarity / scale, dim=-1)  # here we do just a standard matrix multiplication out = einsum(attention, value, "b h n s, b h s d -> b h n d")  # finally, just reshape back to the (batch_size, seq_len, num_kv_heads, hidden_dim) out = rearrange(out, "b h n d -> b n h d") print(out.shape) # torch.Size([1, 256, 2, 64])