Add support for PyTorch style printing of output tensors

examples/apple/mps/executor_runner/mps_executor_runner.mm

@@ -17,6 +17,7 @@
 #import <MetalPerformanceShadersGraph/MetalPerformanceShadersGraph.h>
 
 #include <memory>
+#include <iostream>
 
 #include <gflags/gflags.h>
 
@@ -31,6 +32,7 @@
 #include <executorch/extension/data_loader/buffer_data_loader.h>
 #include <executorch/runtime/core/result.h>
 #include <executorch/runtime/platform/runtime.h>
+#include <executorch/extension/evalue_util/print_evalue.h>
 
 #include <chrono>
 using namespace std::chrono;
@@ -440,15 +442,10 @@ MemoryManager memory_manager(
   std::vector<EValue> outputs(method->outputs_size());
   status = method->get_outputs(outputs.data(), outputs.size());
   ET_CHECK(status == Error::Ok);
-  for (EValue& output : outputs) {
-    // TODO(T159700776): This assumes that all outputs are fp32 tensors. Add
-    // support for other EValues and Tensor dtypes, and print tensors in a more
-    // readable way.
-    auto output_tensor = output.toTensor();
-    auto data_output = output_tensor.const_data_ptr<float>();
-    for (size_t j = 0; j < output_tensor.numel(); ++j) {
-      ET_LOG(Info, "%f", data_output[j]);
-    }
+  // Print the first and last 100 elements of long lists of scalars.
+  std::cout << torch::executor::util::evalue_edge_items(100);
+  for (int i = 0; i < outputs.size(); ++i) {
+    std::cout << "Output " << i << ": " << outputs[i] << std::endl;
   }
 
   // Dump the profiling data to the specified file.

examples/apple/mps/executor_runner/targets.bzl

@@ -15,6 +15,7 @@ def define_common_targets():
             deps = [
                 "//executorch/backends/apple/mps/runtime:MPSBackend",
                 "//executorch/runtime/executor:program",
+                "//executorch/extension/evalue_util:print_evalue",
                 "//executorch/extension/data_loader:file_data_loader",
                 "//executorch/kernels/portable:generated_lib_all_ops",
                 "//executorch/extension/data_loader:file_data_loader",