Merge branch 'v0.11.0-dev-yql' into 'v0.11.0-dev'

解决gptq的不能开启graph的问题

See merge request dcutoolkit/deeplearing/vllm!304

vllm/_custom_ops.py

@@ -765,22 +765,33 @@ def gptq_gemm(a: torch.Tensor, b_q_weight: torch.Tensor,
               b_gptq_qzeros: torch.Tensor, b_gptq_scales: torch.Tensor,
               b_g_idx: torch.Tensor, use_exllama: bool,
               bit: int) -> torch.Tensor:
+    #return quant_ops.gptq_gemm_(a, b_q_weight, b_gptq_qzeros, b_gptq_scales,
+    #                              b_g_idx, use_exllama, bit)
+    return torch.ops.vllm.gptq_gemm_(a, b_q_weight, b_gptq_qzeros, b_gptq_scales,
+                                  b_g_idx, use_exllama, bit)
+def gptq_gemm_(a: torch.Tensor, b_q_weight: torch.Tensor,
+              b_gptq_qzeros: torch.Tensor, b_gptq_scales: torch.Tensor,
+              b_g_idx: torch.Tensor, use_exllama: bool,
+              bit: int) -> torch.Tensor:
     return quant_ops.gptq_gemm(a, b_q_weight, b_gptq_qzeros, b_gptq_scales,
                                   b_g_idx, use_exllama, bit)
-    # return torch.ops._C.gptq_gemm(a, b_q_weight, b_gptq_qzeros, b_gptq_scales,
-    #                               b_g_idx, use_exllama, bit)
+def gptq_gemm_fake_(a: torch.Tensor, b_q_weight: torch.Tensor,
+              b_gptq_qzeros: torch.Tensor, b_gptq_scales: torch.Tensor,
+              b_g_idx: torch.Tensor, use_exllama: bool,
+              bit: int) -> torch.Tensor:
+    return torch.empty((a.shape[0], b_gptq_scales.shape[1]), dtype=a.dtype, device=a.device)
 
 
-if hasattr(torch.ops._C, "gptq_gemm"):
+# if hasattr(torch.ops._C, "gptq_gemm"):
 
-    @register_fake("_C::gptq_gemm")
-    def _gptq_gemm_fake(a: torch.Tensor, b_q_weight: torch.Tensor,
-                        b_gptq_qzeros: torch.Tensor,
-                        b_gptq_scales: torch.Tensor, b_g_idx: torch.Tensor,
-                        use_exllama: bool, bit: int) -> torch.Tensor:
-        return torch.empty((a.size(0), b_q_weight.size(1)),
-                           dtype=a.dtype,
-                           device=a.device)
+#     @register_fake("_C::gptq_gemm")
+#     def _gptq_gemm_fake(a: torch.Tensor, b_q_weight: torch.Tensor,
+#                         b_gptq_qzeros: torch.Tensor,
+#                         b_gptq_scales: torch.Tensor, b_g_idx: torch.Tensor,
+#                         use_exllama: bool, bit: int) -> torch.Tensor:
+#         return torch.empty((a.size(0), b_q_weight.size(1)),
+#                            dtype=a.dtype,
+#                            device=a.device)
 
 
 def gptq_shuffle(q_weight: torch.Tensor, q_perm: torch.Tensor,
@@ -2601,3 +2612,10 @@ direct_register_custom_op(
     mutates_args=[],
     fake_impl=awq_gemm_fake,
 )
+
+direct_register_custom_op(
+    op_name="gptq_gemm_",
+    op_func=gptq_gemm_,
+    mutates_args=[],
+    fake_impl=gptq_gemm_fake_,
+)