Upload modeling_opt.py

modeling_opt.py

@@ -201,10 +201,10 @@ class OPTAttention(nn.Module):
         self.scaling = self.head_dim**-0.5
         self.is_decoder = is_decoder
 
-        self.k_proj = nn.Linear(embed_dim, embed_dim, bias=bias)
-        self.v_proj = nn.Linear(embed_dim, embed_dim, bias=bias)
-        self.q_proj = nn.Linear(embed_dim, embed_dim, bias=bias)
-        self.out_proj = nn.Linear(embed_dim, embed_dim, bias=bias)
+        self.k_proj = nn.Linear(self.embed_dim, self.embed_dim, bias=bias)
+        self.v_proj = nn.Linear(self.embed_dim, self.embed_dim, bias=bias)
+        self.q_proj = nn.Linear(self.embed_dim, self.embed_dim, bias=bias)
+        self.out_proj = nn.Linear(self.embed_dim, self.embed_dim, bias=bias)
 
         # YB: capture the input and output of the softmax
         self.attn_scores = nn.Identity()  # before attention mask