[https://nvbugs/5572320][fix] Ported test_ad_trtllm_bench.py from main (#8671)

Signed-off-by: Eran Geva <[email protected]>

Author: MrGeva

tests/integration/test_lists/test-db/l0_a30.yml

@@ -19,7 +19,7 @@ l0_a30:
   - unittest/_torch/modeling -k "modeling_qwen"
   - unittest/_torch/modeling -k "modeling_qwen_moe"
   - unittest/_torch/modeling -k "modeling_out_of_tree"
-  - unittest/_torch/auto_deploy/unit/singlegpu -k "not test_trtllm_bench_backend_comparison"
+  - unittest/_torch/auto_deploy/unit/singlegpu
   - unittest/_torch/sampler/test_beam_search.py
   - test_e2e.py::test_openai_completions_with_logit_bias[torch_sampler]
   - test_e2e.py::test_openai_chat_with_logit_bias[torch_sampler]

tests/integration/test_lists/test-db/l0_b200.yml

@@ -74,7 +74,7 @@ l0_b200:
   - unittest/_torch/modeling -k "modeling_mixtral"
   - unittest/_torch/modeling -k "modeling_gpt_oss"
   - unittest/_torch/modeling/test_modeling_exaone4.py::TestEXAONE4::test_llm_load_1_FP8
-  - unittest/_torch/auto_deploy/unit/singlegpu -k "not test_trtllm_bench_backend_comparison"
+  - unittest/_torch/auto_deploy/unit/singlegpu
 - condition:
     ranges:
       system_gpu_count:

tests/integration/test_lists/test-db/l0_h100.yml

@@ -31,7 +31,6 @@ l0_h100:
   - unittest/_torch/modeling -k "modeling_nemotron"
   - unittest/_torch/modeling -k "modeling_gemma3"
   - unittest/_torch/modeling -k "modeling_gpt_oss"
-  - unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py::test_trtllm_bench_backend_comparison
   - unittest/disaggregated/test_disagg_utils.py
   - unittest/disaggregated/test_router.py
   - unittest/disaggregated/test_remoteDictionary.py

tests/unittest/_torch/auto_deploy/_utils_test/_model_test_utils.py

@@ -446,6 +446,12 @@ def apply_rotary_pos_emb_ds(q, k, cos, sin, position_ids, unsqueeze_dim=1):
             "vision_config": {"num_hidden_layers": 2},
         },
     },
+    "TinyLlama/TinyLlama-1.1B-Chat-v1.0": {
+        "llm_models_subdir": "llama-models-v2/TinyLlama-1.1B-Chat-v1.0",
+        "model_kwargs": {
+            "num_hidden_layers": 2,
+        },
+    },
 }