TensorFlower Gardener <******@****.***> 1672352909 -0800

---
name: TensorFlow Lite Converter Issue
about: Use this template for reporting issues during model conversion to TFLite
labels: 'TFLiteConverter'

---

### 1. System information

- OS Platform and Distribution (e.g., Linux Ubuntu 16.04):
- TensorFlow installation (pip package or built from source):
- TensorFlow library (version, if pip package or github SHA, if built from source):

### 2. Code

A. Unique TensorFlower <******@****.***> 1655263956 -0700

A. Unique TensorFlower <******@****.***> 1655263956 -0700

    case tflite::TensorType_FLOAT16:
      return builder.getF16Type();
    case tflite::TensorType_BFLOAT16:
      return builder.getBF16Type();
    case tflite::TensorType_FLOAT32:
      return builder.getF32Type();
    case tflite::TensorType_FLOAT64:
      return builder.getF64Type();
    case tflite::TensorType_INT32:
      return builder.getIntegerType(32);
    case tflite::TensorType_UINT16:

                             const tflite::Model* input_model,
                             BufferType quant_type,
                             bool use_updated_hybrid_scheme) {
  tflite::TensorType inference_type;
  switch (quant_type) {
    case BufferType::QUANTIZED_FLOAT16:
      inference_type = tflite::TensorType_FLOAT16;
      break;
    default:
      inference_type = tflite::TensorType_INT8;
  }

namespace mlir {
class Builder;
}  // namespace mlir

namespace tflite {
// Convert the MLIR type to the corresponding TFLite tensor.
tflite::TensorType ConvertTypeToTensorType(mlir::Type type);

// Convert the scalar type of a TFlite tensor to the corresponding MLIR type.
mlir::Type ConvertElementType(tflite::TensorType type, mlir::Builder builder);

// Convert the scalar type of a TFLite tensor to the corresponding
// Tensorflow type

// A list of filters for TAC users to run ops/functions on ML hardwares. The
// intuition is that, for ops/functions that can be run on ML hardware (e.g.
// EdgeTPU) and TFLite CPU, TAC users give a hint that they're more performant
// to run on TFLite CPU. These filters give the TAC users freedom to specify the
// parts that they want to use other hardware to accelerate.
message TacFilters {

    return 1;

  // Create TFLite interpreter & invoke converted program.
  std::unique_ptr<tflite::FlatBufferModel> model =
      tflite::FlatBufferModel::BuildFromBuffer(serialized_flatbuffer.c_str(),
                                               serialized_flatbuffer.size());
  tflite::ops::builtin::BuiltinOpResolver builtins;
  std::unique_ptr<tflite::Interpreter> interpreter;

#include "tensorflow/lite/model.h"

using llvm::cl::opt;

// RUN: flatbuffer_translate -mlir-to-tflite-flatbuffer %s.mlir -o - \
// RUN:   | %p/importer_test_min_max - \
// RUN:   | flatbuffer_translate --tflite-flatbuffer-to-mlir - -o - \
// RUN:   | FileCheck %s

// RUN: flatbuffer_translate -mlir-to-tflite-flatbuffer %s.mlir -o - \
// RUN:   | %p/importer_test_min_max - \
// RUN:   | flatbuffer_to_string - \