[UX] Add vLLM model inspection view (#29450)

Signed-off-by: mgoin <mgoin64@gmail.com>

vllm/envs.py

@@ -250,6 +250,7 @@ if TYPE_CHECKING:
     VLLM_SHARED_EXPERTS_STREAM_TOKEN_THRESHOLD: int = 256
     VLLM_COMPILE_CACHE_SAVE_FORMAT: Literal["binary", "unpacked"] = "binary"
     VLLM_USE_V2_MODEL_RUNNER: bool = False
+    VLLM_LOG_MODEL_INSPECTION: bool = False
     VLLM_DEBUG_MFU_METRICS: bool = False
 
 
@@ -1595,6 +1596,12 @@ environment_variables: dict[str, Callable[[], Any]] = {
     "VLLM_USE_V2_MODEL_RUNNER": lambda: bool(
         int(os.getenv("VLLM_USE_V2_MODEL_RUNNER", "0"))
     ),
+    # Log model inspection after loading.
+    # If enabled, logs a transformers-style hierarchical view of the model
+    # with quantization methods and attention backends.
+    "VLLM_LOG_MODEL_INSPECTION": lambda: bool(
+        int(os.getenv("VLLM_LOG_MODEL_INSPECTION", "0"))
+    ),
     # Debug logging for --enable-mfu-metrics
     "VLLM_DEBUG_MFU_METRICS": lambda: bool(
         int(os.getenv("VLLM_DEBUG_MFU_METRICS", "0"))