code review feedback

src/axolotl/monkeypatch/moe/moe.py

@@ -25,24 +25,33 @@ class SparseMoeBlock(nn.Module):
         w1s, w3s = torch.split(torch.unbind(self.experts.experts.weight, dim=0), 2, dim=1)
         w2s = torch.unbind(self.experts.output_experts.weight, dim=0)
 
-        # Recreate the MoE class with original weights
+        # Recreate the structure of the original MixtralSparseMoeBlock
-        experts = []
+        original_moe = nn.Module()
-        for i in range(self.num_experts):
+        original_moe.hidden_dim = self.hidden_dim
-            expert = nn.Sequential(
+        original_moe.ffn_dim = self.ffn_dim
-                nn.Linear(self.hidden_dim, 2 * self.ffn_dim, bias=False),
+        original_moe.num_experts = self.num_experts
-                self.experts.activation,
+        original_moe.top_k = self.top_k
-                nn.Linear(self.ffn_dim, self.hidden_dim, bias=False),
-            )
-            expert[0].weight.data = torch.cat([w1s[i], w3s[i]], dim=0)
-            expert[2].weight.data = w2s[i]
-            experts.append(expert)
 
-        # Create a new MoE module with the recreated experts
+        # Recreate the gating module
-        moe = nn.ModuleList(experts)
+        original_moe.gate = nn.Linear(self.hidden_dim, self.num_experts, bias=False)
+        original_moe.gate.weight.data = self.gate.weight.data
 
-        # Replace the fused experts with the recreated MoE module
+        # Recreate the experts as a ModuleList
-        setattr(model, name.replace("experts", "moe"), moe)
+        original_moe.experts = nn.ModuleList()
-        delattr(model, name)
+        for expert_idx in range(self.num_experts):
+            expert = nn.Module()
+            expert.w1 = nn.Linear(self.hidden_dim, 2 * self.ffn_dim, bias=False)
+            expert.w2 = nn.Linear(self.ffn_dim, self.hidden_dim, bias=False)
+            expert.w3 = nn.Linear(self.hidden_dim, 2 * self.ffn_dim, bias=False)
+            expert.act_fn = self.experts.activation
+
+            expert.w1.weight.data = torch.cat([w1s[expert_idx], w3s[expert_idx]], dim=0)
+            expert.w2.weight.data = w2s[expert_idx]
+
+            original_moe.experts.append(expert)
+
+        # Replace the SparseMoeBlock with the recreated MixtralSparseMoeBlock structure
+        setattr(model, name, original_moe)
 
     def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         batch_size, sequence_length, hidden_dim = hidden_states.shape