Merge pull request #2319 from ftgreat/main

Add patch_mixtral_replace_moe_impl for full training Mitral using DeepSpeed Zero3

src/llmtuner/model/patcher.py

@@ -268,6 +268,43 @@ def patch_config(
     _configure_quantization(config, tokenizer, model_args, config_kwargs)
 
 
+def patch_mixtral_replace_moe_impl() -> None:
+    import torch.nn.functional as F
+    def mlp_forward(self, hidden_states):
+        current_hidden_states = self.act_fn(self.w1(hidden_states)) * self.w3(hidden_states)
+        current_hidden_states = self.w2(current_hidden_states)
+        return current_hidden_states
+
+    ## Ref. https://huggingface.co/deepseek-ai/deepseek-moe-16b-base/blob/main/modeling_deepseek.py
+    def moe_forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        batch_size, sequence_length, hidden_dim = hidden_states.shape
+        hidden_states = hidden_states.view(-1, hidden_dim)
+        # router_logits: (batch * sequence_length, n_experts)
+        router_logits = self.gate(hidden_states)
+
+        routing_weights = F.softmax(router_logits, dim=1, dtype=torch.float)
+        topk_weight, topk_idx = torch.topk(routing_weights, self.top_k, dim=-1, sorted=False)
+        topk_weight /= topk_weight.sum(dim=-1, keepdim=True)
+        # we cast back to the input dtype
+        topk_weight = topk_weight.to(hidden_states.dtype)
+
+        hidden_states = hidden_states.repeat_interleave(self.top_k, dim=0)
+        y = torch.empty_like(hidden_states)
+        flat_topk_idx = topk_idx.view(-1)
+        for i in range(self.num_experts):
+            expert = self.experts[i]
+            y[flat_topk_idx == i] = expert(hidden_states[flat_topk_idx == i])
+        y = (y.view(*topk_weight.shape, -1) * topk_weight.unsqueeze(-1)).sum(dim=1)
+        final_hidden_states = y.reshape(batch_size, sequence_length, hidden_dim)
+        return final_hidden_states, router_logits
+
+    from transformers.models.mixtral.modeling_mixtral import MixtralSparseMoeBlock
+    from transformers.models.mixtral.modeling_mixtral import MixtralBLockSparseTop2MLP
+
+    MixtralBLockSparseTop2MLP.forward = mlp_forward
+    MixtralSparseMoeBlock.forward = moe_forward
+
+
 def patch_model(
     model: "PreTrainedModel", tokenizer: "PreTrainedTokenizer", model_args: "ModelArguments", is_trainable: bool
 ) -> None:
@@ -290,6 +327,9 @@ def patch_model(
         from transformers.models.mixtral.modeling_mixtral import MixtralSparseMoeBlock
         set_z3_leaf_modules(model, [MixtralSparseMoeBlock])
 
+        if is_trainable:
+            patch_mixtral_replace_moe_impl()
+
 
 def patch_valuehead_model(model: "AutoModelForCausalLMWithValueHead") -> None:
     def tie_weights(self: "AutoModelForCausalLMWithValueHead") -> None: