Fix vllm tests

Signed-off-by: Onur Yilmaz <oyilmaz@nvidia.com>

Author: oyilmaz-nvidia

docs/llm/nemo_2/optimized/vllm.md

@@ -81,7 +81,6 @@ After executing the script, it will export the model to vLLM and then initiate t
    - ``--swap_space``: Size (GiB) of CPU memory per GPU to use as swap space. Default is 4.
    - ``--cpu_offload_gb``: Size (GiB) of CPU memory to use for offloading model weights. Default is 0.
    - ``--enforce_eager``: Whether to enforce eager execution. Default is False.
-   - ``--max_seq_len_to_capture``: Maximum sequence length covered by CUDA graphs. Default is 8192.
    - ``--triton_model_name``: Name for the service/model on Triton. (Required)
    - ``--triton_model_version``: Version for the service/model. Default is 1.
    - ``--triton_port``: Port for the Triton server to listen for requests. Default is 8000.

tests/unit_tests/export/test_vllm_exporter.py

@@ -70,7 +70,6 @@ def test_export(exporter, mock_llm):
         swap_space=4,
         cpu_offload_gb=0,
         enforce_eager=False,
-        max_seq_len_to_capture=8192,
         task="auto",
     )
 
@@ -96,7 +95,6 @@ def test_export_with_lora(exporter, mock_llm):
         swap_space=4,
         cpu_offload_gb=0,
         enforce_eager=False,
-        max_seq_len_to_capture=8192,
         task="auto",
     )
 
@@ -128,7 +126,6 @@ def test_export_with_custom_params(exporter, mock_llm):
         swap_space=4,
         cpu_offload_gb=0,
         enforce_eager=False,
-        max_seq_len_to_capture=8192,
         task="auto",
     )
 
@@ -844,7 +841,6 @@ def test_export_megatron_bridge_with_all_vllm_params(exporter, mock_llm):
             swap_space=8,
             cpu_offload_gb=2,
             enforce_eager=True,
-            max_seq_len_to_capture=4096,
             task="generate",
         )
 
@@ -863,5 +859,4 @@ def test_export_megatron_bridge_with_all_vllm_params(exporter, mock_llm):
         assert call_kwargs["swap_space"] == 8
         assert call_kwargs["cpu_offload_gb"] == 2
         assert call_kwargs["enforce_eager"] is True
-        assert call_kwargs["max_seq_len_to_capture"] == 4096
         assert call_kwargs["task"] == "generate"