Update

Signed-off-by: Chenfei Zhang <[email protected]>

Author: chenfeiz0326

tensorrt_llm/executor/executor.py

@@ -135,8 +135,6 @@ def generate_async(
             postproc_params.postproc_args.num_prompt_tokens = len(
                 prompt_token_ids)
 
-        print(f"[CF][generate_async] sampling_params is ")
-        print(sampling_params)
         request = GenerationRequest(
             prompt_token_ids,
             sampling_params=sampling_params,

tests/integration/defs/accuracy/test_llm_api_pytorch.py

@@ -491,7 +491,6 @@ def test_nvfp4_tp4(self):
         kv_cache_config = KvCacheConfig(free_gpu_memory_fraction=0.5)
         with LLM(model_path,
                  tensor_parallel_size=4,
-                 max_seq_len=8192,
                  max_batch_size=32,
                  kv_cache_config=kv_cache_config) as llm:
             assert llm.args.quant_config.quant_algo == QuantAlgo.NVFP4