Add hf_T5_generate to inline_inbuilt_nn_modules (#131804)

Summary:
Fixes pytorch/pytorch#121989

We are turning on the flag by default in another PR. But that PR can go
through reverts. So, forcibly adding the benchmark to prevent dashboard
fluctuation in case of reverts.

X-link: pytorch/pytorch#131804
Approved by: https://github.com/yanboliang, https://github.com/shunting314
ghstack dependencies: #131795, #131801

Reviewed By: clee2000

Differential Revision: D60269692

Pulled By: anijain2305

fbshipit-source-id: d014844a90a621eb99ba511200b21c930a9f5638

userbenchmark/dynamo/dynamobench/torchbench.py

@@ -242,6 +242,7 @@ def inline_inbuilt_nn_modules_models(self):
             "detectron2_maskrcnn_r_101_fpn",
             "vision_maskrcnn",
             "doctr_reco_predictor",
+            "hf_T5_generate",
         }
 
     def load_model(