minor changes

Signed-off-by: leslie-fang25 <[email protected]>

Author: leslie-fang25

tensorrt_llm/executor/worker.py

@@ -788,7 +788,6 @@ def notify_proxy_threads_to_quit():
                         "green")
 
     try:
-        print("---- worker_cls is: {}".format(worker_cls), flush=True)
         worker: GenerationExecutorWorker = worker_cls(
             engine,
             executor_config,

tensorrt_llm/llmapi/llm.py

@@ -37,7 +37,7 @@
 from .llm_utils import (CachedModelLoader, KvCacheRetentionConfig,
                         LlmBuildStats, ModelLoader, _ModelRuntimeContext)
 from .mpi_session import MpiPoolSession, external_mpi_comm_available
-from .tokenizer import (TokenizerBase, _xgrammar_tokenizer_info)
+from .tokenizer import TokenizerBase, _xgrammar_tokenizer_info
 # TODO[chunweiy]: move the following symbols back to utils scope, and remove the following import
 from .utils import (append_docstring, exception_handler, get_device_count,
                     print_colored_debug, set_api_status)
@@ -959,27 +959,8 @@ def _build_model(self):
 
         assert isinstance(self.args, TorchLlmArgs)
 
-        # self._executor_config = tllm.ExecutorConfig(
-        #     max_beam_width=self.args.max_beam_width,
-        #     scheduler_config=PybindMirror.maybe_to_pybind(
-        #         self.args.scheduler_config),
-        #     max_batch_size=self.args.max_batch_size,
-        #     max_num_tokens=self.args.max_num_tokens,
-        #     gather_generation_logits=self.args.gather_generation_logits,
-        #     fail_fast_on_attention_window_too_large=getattr(
-        #         self.args, 'fail_fast_on_attention_window_too_large', False),
-        #     **kwargs)
-
-        # self._executor_config = self.args.get_executor_config(self._hf_model_dir)
-
         # TODO: revisit gather_context_logits
         return_logits = self.args.gather_generation_logits
-
-        print("---- self._executor_cls is: {}".format(self._executor_cls),
-              flush=True)
-        print("---- self._engine_dir is: {}".format(self._engine_dir),
-              flush=True)
-
         self._executor = self._executor_cls.create(
             self._engine_dir,
             executor_config=None,