[fix]修复不开mtp精度异常问题

vllm/envs.py

@@ -1188,7 +1188,7 @@ environment_variables: dict[str, Callable[[], Any]] = {
                  ("true", "1")),
 
     "VLLM_USE_ZERO_MTP":
-        lambda: (os.getenv('VLLM_USE_ZERO_MTP', '0').lower() in
+        lambda: (os.getenv('VLLM_USE_ZERO_MTP', '1').lower() in
                  ("true", "1")),
 
     # vllm will use 1-24... (not only 1 2 4 8 16 24)

vllm/v1/worker/gpu_model_runner.py

@@ -3315,11 +3315,11 @@ class GPUModelRunnerMTP(GPUModelRunnerBase):
         # Get the valid generated tokens.
         sampled_token_ids = sampler_output.sampled_token_ids
         max_gen_len = sampled_token_ids.shape[-1]
-        sampled_token_ids_cpu = sampled_token_ids.to('cpu', non_blocking=True)
         if not self.speculative_config:
             # Speculative decoding is not enabled.
             spec_token_ids = None
         else:
+            sampled_token_ids_cpu = sampled_token_ids.to('cpu', non_blocking=True)
             self.spec_sampler_event.record()
             mask = (sampled_token_ids == -1)
             mask_int = mask.int()
@@ -3338,7 +3338,7 @@ class GPUModelRunnerMTP(GPUModelRunnerBase):
             )
         if max_gen_len == 1:
             # No spec decode tokens.
-            valid_sampled_token_ids = sampled_token_ids_cpu.tolist()
+            valid_sampled_token_ids = sampled_token_ids.tolist()
         else:
             # Includes spec decode tokens.
             self.spec_sampler_event.synchronize()