Fail the model evaluation if there's no successful TFLite interpreter invocations in all attempts.

PiperOrigin-RevId: 258908895

tensorflow/lite/tools/accuracy/ilsvrc/imagenet_accuracy_eval.cc

@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include <cstdlib>
 #include <iomanip>
 #include <memory>
 #include <mutex>  // NOLINT(build/c++11)
@@ -178,17 +179,17 @@ int Main(int argc, char* argv[]) {
   std::unique_ptr<ImagenetModelEvaluator> evaluator;
   if (output_file_path.empty()) {
     LOG(ERROR) << "Invalid output file path.";
-    return 0;
+    return EXIT_FAILURE;
   }
 
   if (num_threads <= 0) {
     LOG(ERROR) << "Invalid number of threads.";
-    return 0;
+    return EXIT_FAILURE;
   }
 
   if (ImagenetModelEvaluator::Create(argc, argv, num_threads, &evaluator) !=
       kTfLiteOk)
-    return 0;
+    return EXIT_FAILURE;
 
   std::ofstream output_stream(output_file_path, std::ios::out);
   if (!output_stream) {
@@ -210,7 +211,10 @@ int Main(int argc, char* argv[]) {
       absl::make_unique<CSVWriter>(columns, &output_stream));
   evaluator->AddObserver(&results_writer);
   LOG(ERROR) << "Starting evaluation with: " << num_threads << " threads.";
-  evaluator->EvaluateModel();
+  if (evaluator->EvaluateModel() != kTfLiteOk) {
+    LOG(ERROR) << "Failed to evaluate the model!";
+    return EXIT_FAILURE;
+  }
 
   if (!proto_output_file_path.empty()) {
     std::ofstream proto_out_file(proto_output_file_path,
@@ -220,7 +224,7 @@ int Main(int argc, char* argv[]) {
     proto_out_file.close();
   }
 
-  return 0;
+  return EXIT_SUCCESS;
 }
 
 }  // namespace metrics

tensorflow/lite/tools/accuracy/ilsvrc/imagenet_model_evaluator.cc

@@ -263,7 +263,7 @@ TfLiteStatus ImagenetModelEvaluator::EvaluateModel() const {
                  &all_okay]() {
       if (EvaluateModelForShard(shard_id, image_label, model_labels, params_,
                                 &observer, params_.num_ranks) != kTfLiteOk) {
-        all_okay = all_okay && false;
+        all_okay = false;
       }
     };
     thread_pool.push_back(std::thread(func));
@@ -274,7 +274,7 @@ TfLiteStatus ImagenetModelEvaluator::EvaluateModel() const {
     thread.join();
   }
 
-  return kTfLiteOk;
+  return all_okay ? kTfLiteOk : kTfLiteError;
 }
 
 }  // namespace metrics

tensorflow/lite/tools/evaluation/stages/tflite_inference_stage.cc

@@ -140,7 +140,10 @@ TfLiteStatus TfliteInferenceStage::Run() {
   auto& params = config_.specification().tflite_inference_params();
   for (int i = 0; i < params.invocations_per_run(); ++i) {
     int64_t start_us = profiling::time::NowMicros();
-    interpreter_->Invoke();
+    if (interpreter_->Invoke() != kTfLiteOk) {
+      LOG(ERROR) << "TFLite interpreter failed to invoke at run " << i;
+      return kTfLiteError;
+    }
     latency_stats_.UpdateStat(profiling::time::NowMicros() - start_us);
   }