remove timing

intel-analytics · Dec 3, 2024 · b62ae49 · b62ae49
1 parent cb8a852
commit b62ae49
Showing 1 changed file with 0 additions and 4 deletions.
diff --git a/python/llm/src/ipex_llm/transformers/npu_models/convert.py b/python/llm/src/ipex_llm/transformers/npu_models/convert.py
@@ -438,7 +438,6 @@ def causal_lm_forward(
     output_hidden_states: Optional[bool] = None,
     return_dict: Optional[bool] = None,
 ) -> Union[Tuple, CausalLMOutputWithPast]:
-    start = time.perf_counter()
     from .npu_llm_cpp import run_prefill_with_logits, run_decode_with_logits
     if isinstance(input_ids[0], torch.Tensor):
         input_list = input_ids[0].flatten().tolist()
@@ -449,9 +448,6 @@ def causal_lm_forward(
         logits = run_prefill_with_logits(self.model_ptr, input_list, self.logits_buffer, self.vocab_size)
     else:
         logits = run_decode_with_logits(self.model_ptr, input_list[0], self.logits_buffer, self.vocab_size)
-    end = time.perf_counter()
-    overall = (end - start) * 1000
-    print("Overall time: ", overall)
 
     return CausalLMOutputWithPast(
         loss=None,