bacow/happy-llm
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update 第二章 Transformer架构.md

Browse Source
This commit is contained in:
Logan Zou
2025-06-23 10:50:16 +08:00
committed by GitHub
parent bd3fb6cf48
commit 185a212acd
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
docs/chapter2/第二章 Transformer架构.md
View File

@@ -266,7 +266,7 @@ class MultiHeadAttention(nn.Module):
self.wq = nn.Linear(args.dim, args.n_heads * self.head_dim, bias=False)
self.wk = nn.Linear(args.dim, args.n_heads * self.head_dim, bias=False)
self.wv = nn.Linear(args.dim, args.n_heads * self.head_dim, bias=False)
# 输出权重矩阵，维度为 n_embd x n_embd（head_dim = n_embeds / n_heads）
# 输出权重矩阵，维度为 dim x n_embd（head_dim = n_embeds / n_heads）
self.wo = nn.Linear(args.n_heads * self.head_dim, args.dim, bias=False)
# 注意力的 dropout
self.attn_dropout = nn.Dropout(args.dropout)