File tree Expand file tree Collapse file tree 2 files changed +2
-15
lines changed Expand file tree Collapse file tree 2 files changed +2
-15
lines changed Original file line number Diff line number Diff line change 1414from  tensorrt_llm ._utils  import  get_sm_version 
1515from  tensorrt_llm .bindings .executor  import  (CapacitySchedulerPolicy ,
1616                                            ContextChunkingPolicy ,
17-                                             ExecutorConfig ,
18-                                             LogitsPostProcessorConfig ,
19-                                             ParallelConfig )
17+                                             ExecutorConfig )
2018from  tensorrt_llm .bindings .internal .batch_manager  import  ContextChunkingConfig 
2119from  tensorrt_llm .llmapi .llm_args  import  KvCacheConnectorConfig , TorchLlmArgs 
2220from  tensorrt_llm .llmapi .tokenizer  import  TokenizerBase 
@@ -217,14 +215,9 @@ def create_py_executor(
217215    tokenizer : Optional [TokenizerBase ] =  None ,
218216    lora_config : Optional [LoraConfig ] =  None ,
219217    kv_connector_config : Optional [KvCacheConnectorConfig ] =  None ,
220-     logits_post_processor_config : Optional [LogitsPostProcessorConfig ] =  None ,
221-     parallel_config : Optional [ParallelConfig ] =  None ,
222218) ->  PyExecutor :
223219
224220    executor_config  =  llm_args .get_executor_config (checkpoint_dir , tokenizer )
225-     executor_config .logits_post_processor_config  =  logits_post_processor_config 
226-     executor_config .parallel_config  =  parallel_config 
227- 
228221    garbage_collection_gen0_threshold  =  llm_args .garbage_collection_gen0_threshold 
229222
230223    _mangle_executor_config (executor_config )
Original file line number Diff line number Diff line change @@ -113,6 +113,7 @@ def _create_py_executor():
113113            assert  hasattr (
114114                self .llm_args , "backend" 
115115            ), "llm_args should be with backend in _create_py_executor" 
116+             _  =  _get_comm_ranks_device_id ()
116117            if  self .llm_args .backend  ==  "pytorch" :
117118                from  tensorrt_llm ._torch .pyexecutor .py_executor_creator  import  \
118119                    create_py_executor 
@@ -122,13 +123,6 @@ def _create_py_executor():
122123                args ["tokenizer" ] =  tokenizer 
123124                args ["lora_config" ] =  lora_config 
124125                args ["kv_connector_config" ] =  kv_connector_config 
125-                 args [
126-                     "logits_post_processor_config" ] =  tllm .LogitsPostProcessorConfig (
127-                         processor_batched = batched_logits_processor ,
128-                         replicate = False )
129-                 comm_ranks , device_ids  =  _get_comm_ranks_device_id ()
130-                 args ["parallel_config" ] =  tllm .ParallelConfig (
131-                     participant_ids = comm_ranks , device_ids = device_ids )
132126            elif  self .llm_args .backend  ==  "_autodeploy" :
133127                from  tensorrt_llm ._torch .auto_deploy .llm_args  import  \
134128                    LlmArgs  as  ADLlmArgs 
    
 
   
 
     
   
   
          
     
  
    
     
 
    
      
     
 
     
    You can’t perform that action at this time.
  
 
    
  
     
    
      
        
     
 
       
      
     
   
 
    
    
  
 
  
 
     
    
0 commit comments