add fuse_attention_ffn support for qwen (#8526)

paddlenlp/transformers/qwen/configuration.py

@@ -43,6 +43,8 @@ def __init__(
         use_flash_attention=False,
         use_fused_rms_norm=False,
         use_fused_rope=False,
+        fuse_attention_ffn=False,
+        sequence_parallel=False,
         intermediate_size=22016,
         tensor_parallel_output=True,
         no_bias=True,
@@ -77,6 +79,8 @@ def __init__(
         self.use_flash_attention = use_flash_attention
         self.use_fused_rms_norm = use_fused_rms_norm
         self.use_fused_rope = use_fused_rope
+        self.fuse_attention_ffn = fuse_attention_ffn
+        self.sequence_parallel = sequence_parallel
         self.no_bias = no_bias
 
         self.long_sequence_strategy_type = long_sequence_strategy_type

paddlenlp/transformers/qwen/modeling.py

@@ -26,6 +26,16 @@
 from paddle.distributed.fleet.utils import recompute
 from paddle.utils import try_import
 
+try:
+    from paddle.incubate.nn.functional import swiglu
+except ImportError:
+
+    def swiglu(x, y=None):
+        if y is None:
+            x, y = paddle.chunk(x, chunks=2, axis=-1)
+        return F.silu(x) * y
+
+
 from paddlenlp.transformers.long_sequence_strategies import LongSequenceStrategies
 from paddlenlp.transformers.model_outputs import (
     BaseModelOutputWithPast,
@@ -35,6 +45,7 @@
 from paddlenlp.utils.log import logger
 
 from ...utils.converter import StateDictNameMapping, init_name_mappings
+from .. import linear_utils
 from ..model_outputs import ModelOutput
 from .configuration import QWenConfig
 
@@ -329,37 +340,60 @@ class QWenMLP(nn.Layer):
     def __init__(self, config):
         super().__init__()
         ff_dim_in = config.intermediate_size // 2
+        self.fuse_attention_ffn = config.fuse_attention_ffn
+
+        if config.sequence_parallel:
+            ColumnParallelLinear = linear_utils.ColumnSequenceParallelLinear
+            RowParallelLinear = linear_utils.RowSequenceParallelLinear
+        else:
+            ColumnParallelLinear = linear_utils.ColumnParallelLinear
+            RowParallelLinear = linear_utils.RowParallelLinear
+
         if config.tensor_parallel_degree > 1:
-            self.w1 = mpu.ColumnParallelLinear(
-                config.hidden_size,
-                ff_dim_in,
-                gather_output=False,
-                has_bias=False,
-            )
-            self.w2 = mpu.ColumnParallelLinear(
-                config.hidden_size,
-                ff_dim_in,
-                gather_output=False,
-                has_bias=False,
-            )
-            self.c_proj = mpu.RowParallelLinear(
+            if self.fuse_attention_ffn:
+                self.gate_up_fused_proj = ColumnParallelLinear(
+                    config.hidden_size,
+                    ff_dim_in * 2,
+                    gather_output=False,
+                    has_bias=False,
+                )
+            else:
+                self.w1 = ColumnParallelLinear(
+                    config.hidden_size,
+                    ff_dim_in,
+                    gather_output=False,
+                    has_bias=False,
+                )
+                self.w2 = ColumnParallelLinear(
+                    config.hidden_size,
+                    ff_dim_in,
+                    gather_output=False,
+                    has_bias=False,
+                )
+            self.c_proj = RowParallelLinear(
                 ff_dim_in,
                 config.hidden_size,
                 input_is_parallel=True,
                 has_bias=False,
             )
         else:
-            self.w1 = nn.Linear(config.hidden_size, ff_dim_in, bias_attr=not config.no_bias)
-            self.w2 = nn.Linear(config.hidden_size, ff_dim_in, bias_attr=not config.no_bias)
+            if self.fuse_attention_ffn:
+                self.gate_up_fused_proj = nn.Linear(config.hidden_size, ff_dim_in * 2, bias_attr=not config.no_bias)
+            else:
+                self.w1 = nn.Linear(config.hidden_size, ff_dim_in, bias_attr=not config.no_bias)
+                self.w2 = nn.Linear(config.hidden_size, ff_dim_in, bias_attr=not config.no_bias)
             self.c_proj = nn.Linear(ff_dim_in, config.hidden_size, bias_attr=not config.no_bias)
 
     def forward(self, hidden_states):
         # up
-        a1 = self.w1(hidden_states)
-        # gate
-        a2 = self.w2(hidden_states)
-        intermediate_parallel = a1 * F.silu(a2)
-        # down
+        # a1 = self.w1(hidden_states)
+        # # gate
+        # a2 = self.w2(hidden_states)
+        # intermediate_parallel = a1 * F.silu(a2)
+        if self.fuse_attention_ffn:
+            intermediate_parallel = swiglu(self.gate_up_fused_proj(hidden_states))
+        else:
+            intermediate_parallel = swiglu(self.w2(hidden_states), self.w1(hidden_states))
         output = self.c_proj(intermediate_parallel)
         return output