Add chatglm2&3 fuse mlp (#12328)

* add chatglm fuse mlp

python/llm/src/ipex_llm/transformers/convert.py

@@ -1049,6 +1049,12 @@ def _optimize_pre(model, qtype=None):
             model.llm.config.model_type = "llama"
         _optimize_pre(model.llm, qtype=qtype)
         model.llm.config.model_type = "minicpmv"
+    if model.config.architectures is not None \
+       and model.config.architectures[0] in ["ChatGLMModel", "ChatGLMForConditionalGeneration"]:
+        from ipex_llm.transformers.models.chatglm2 import split_mlp
+        if hasattr(model.config, 'padded_vocab_size') and \
+           model.config.padded_vocab_size == 65024:
+                model.apply(split_mlp)
 
     return model
 
@@ -1372,6 +1378,7 @@ def _optimize_post(model, lightweight_bmm=False):
             from ipex_llm.transformers.models.chatglm2 import chatglm_rms_norm_forward
             from ipex_llm.transformers.models.chatglm2 import chatglm2_encoder_forward
             from ipex_llm.transformers.models.chatglm2 import chatglm2_model_forward
+            from ipex_llm.transformers.models.chatglm2 import mlp_forward
             convert_forward(model,
                             module.SelfAttention,
                             chatglm2_attention_forward)
@@ -1384,6 +1391,7 @@ def _optimize_post(model, lightweight_bmm=False):
             convert_forward(model,
                             module.RMSNorm,
                             chatglm_rms_norm_forward)
+            convert_forward(model, module.MLP, mlp_forward)
         elif hasattr(model.config, 'padded_vocab_size') and \
                 model.config.padded_vocab_size == 64896:
             # codegeex-nano

python/llm/src/ipex_llm/transformers/models/chatglm2.py

@@ -26,6 +26,7 @@ from ipex_llm.utils.common.log4Error import invalidInputError
 from ipex_llm.transformers.models.utils import restore_fp8_kv_cache, update_past_key_value
 from ipex_llm.transformers.models.utils import use_quantize_kv_cache, use_sdp, use_sdp_causal
 from ipex_llm.transformers.models.utils import should_use_fuse_rope, apply_rotary_pos_emb
+from ipex_llm.transformers.models.utils import mlp_fusion_check, SILU
 from ipex_llm.transformers.models.utils import use_quantize_kv_cache, use_sdp, \
     use_sdp_causal, should_use_compresskv, is_enough_kv_cache_room_4_36
 from ipex_llm.transformers.kv import DynamicCompressCache, DynamicCompressFp8Cache
@@ -91,7 +92,7 @@ def chatglm2_model_forward(
 
     if use_cache:
         use_compress_kv = should_use_compresskv(input_ids, input_ids.shape[1])
-        use_quantize_kv = use_quantize_kv_cache(self.encoder.layers[0].mlp.dense_h_to_4h,
+        use_quantize_kv = use_quantize_kv_cache(self.encoder.layers[0].mlp.gate_proj,
                                                 input_ids)
         if use_compress_kv and not isinstance(past_key_values,
                                               DynamicCompressCache):
@@ -570,3 +571,46 @@ def codegeex_attention_forward(
     output = self.dense(context_layer)
 
     return output, past_key_value
+
+import torch.nn.functional as F
+
+
+def split_mlp(module: torch.nn.Module):
+    if module.__class__.__name__ == "MLP":
+        gate_weight, up_weight = module.dense_h_to_4h.weight.data.chunk(2, dim=0)
+
+        gate_proj = torch.nn.Linear(0, 0, bias=False)
+        gate_proj.weight = torch.nn.Parameter(gate_weight, requires_grad=False)
+        gate_proj.in_features = gate_weight.size(1)
+        gate_proj.out_features = gate_weight.size(0)
+
+        up_proj = torch.nn.Linear(0, 0, bias=False)
+        up_proj.weight = torch.nn.Parameter(up_weight, requires_grad=False)
+        up_proj.in_features = up_weight.size(1)
+        up_proj.out_features = up_weight.size(0)
+
+        module.gate_proj = gate_proj
+        module.up_proj = up_proj
+
+        module.activation_fn = F.silu
+
+        del module.dense_h_to_4h
+
+
+def mlp_forward(
+    self,
+    hidden_states: torch.FloatTensor
+) -> torch.FloatTensor:
+    x_2d = hidden_states.view(-1, hidden_states.shape[-1])
+    qtype = getattr(self.gate_proj, "qtype", None)
+    if mlp_fusion_check(x_2d, qtype, self.training):
+        x_2d = x_2d.contiguous()
+        import xe_linear
+        return self.dense_4h_to_h(xe_linear.mlp_forward_xpu(
+            x_2d, self.gate_proj.weight.data, self.up_proj.weight.data,
+            x_2d.shape[0], x_2d.shape[1], self.gate_proj.out_features,
+            SILU, qtype
+        ))
+    return self.dense_4h_to_h(
+        self.activation_fn(self.gate_proj(hidden_states)) * self.up_proj(hidden_states)
+    )