[None][chore] Restore asserts in pytorch flow LoRA tests (NVIDIA#8227)

Signed-off-by: Amit Zuker <203509407+amitz-nv@users.noreply.github.com>

Author: amitz-nv

tests/unittest/llmapi/test_llm_pytorch.py

@@ -267,9 +267,13 @@ def llama_7b_lora_from_dir_test_harness(**llm_kwargs) -> None:
         max_lora_rank=8,
         max_loras=2,
         max_cpu_loras=2)
-    llm = LLM(model=f"{llm_models_root()}/llama-models/llama-7b-hf",
-              lora_config=lora_config,
-              **llm_kwargs)
+    llm = LLM(
+        model=f"{llm_models_root()}/llama-models/llama-7b-hf",
+        lora_config=lora_config,
+        # Disable CUDA graph
+        # TODO: remove this once we have a proper fix for CUDA graph in LoRA
+        cuda_graph_config=None,
+        **llm_kwargs)
     try:
         prompts = [
             "美国的首都在哪里? \n答案:",
@@ -285,10 +289,7 @@ def llama_7b_lora_from_dir_test_harness(**llm_kwargs) -> None:
         outputs = llm.generate(prompts,
                                sampling_params,
                                lora_request=lora_request)
-        # TODO: remove this once we have a proper fix for CUDA graph in LoRA
-        # assert similar(outputs[0].outputs[0].text, references[0])
-        print(f"lora output: {outputs[0].outputs[0].text}")
-        print(f"ref output: {references[0]}")
+        assert similar(outputs[0].outputs[0].text, references[0])
     finally:
         llm.shutdown()
 
@@ -304,7 +305,12 @@ def test_llama_7b_lora_default_modules() -> None:
 
     hf_model_dir = f"{llm_models_root()}/llama-models/llama-7b-hf"
 
-    llm = LLM(model=hf_model_dir, lora_config=lora_config)
+    llm = LLM(
+        model=hf_model_dir,
+        lora_config=lora_config,
+        # Disable CUDA graph
+        # TODO: remove this once we have a proper fix for CUDA graph in LoRA
+        cuda_graph_config=None)
 
     hf_lora_dir = f"{llm_models_root()}/llama-models/luotuo-lora-7b-0.1"
     try:
@@ -323,9 +329,7 @@ def test_llama_7b_lora_default_modules() -> None:
                                sampling_params,
                                lora_request=lora_request)
 
-        # assert similar(outputs[0].outputs[0].text, references[0])
-        print(f"lora output: {outputs[0].outputs[0].text}")
-        print(f"ref output: {references[0]}")
+        assert similar(outputs[0].outputs[0].text, references[0])
     finally:
         llm.shutdown()