Targets for Qualcomm llm eval

Differential Revision: D82655293

Pull Request resolved: pytorch#14381

Author: rohansjoshi

examples/qualcomm/oss_scripts/llama/TARGETS

@@ -26,6 +26,16 @@ runtime.python_library(
     ],
 )
 
+runtime.python_library(
+    name = "masking_utils",
+    srcs = [
+        "masking_utils.py",
+    ],
+    deps = [
+        "//caffe2:torch",
+    ],
+)
+
 runtime.python_library(
     name = "decoder_constants",
     srcs = [
@@ -39,6 +49,7 @@ runtime.python_library(
     deps = [
         ":decoder_constants",
         ":decoder_utils",
+        ":masking_utils",
         "//executorch/examples/models/llama:source_transformation",
         "//caffe2:torch",
         "//executorch/backends/qualcomm/partition:partition",
@@ -90,6 +101,7 @@ python_binary(
         "//executorch/examples/qualcomm/oss_scripts/llama:range_setting_pt2e",
         "fbsource//third-party/pypi/lm-eval:lm-eval",
     ],
+    keep_gpu_sections = True,
 )
 
 runtime.command_alias(