//   t[:, 1, :, :] will have scale[1]=2.0, zero_point[0]=2
  //   t[:, 2, :, :] will have scale[2]=3.0, zero_point[0]=3
  quantized_dimension:int;
}

// Sparse tensors.
// We use a modification of the TACO format.
// Reference: http://tensor-compiler.org/kjolstad-oopsla17-tensor-compiler.pdf
//
// To encode a conceptual n-dimensional dense tensor with dims (d0, ..., dn-1),

  //   t[:, 1, :, :] will have scale[1]=2.0, zero_point[0]=2
  //   t[:, 2, :, :] will have scale[2]=3.0, zero_point[0]=3
  quantized_dimension:int;
}

// Sparse tensors.
// We use a modification of the TACO format.
// Reference: http://tensor-compiler.org/kjolstad-oopsla17-tensor-compiler.pdf
//
// To encode a conceptual n-dimensional dense tensor with dims (d0, ..., dn-1),

* _Đường dẫn_ `/items/{item_id}`  có một _tham số query string_ `q`, nó là một tham số tùy chọn kiểu `str`.

### Tài liệu tương tác API

Truy cập <a href="http://127.0.0.1:8000/docs" class="external-link" target="_blank">http://127.0.0.1:8000/docs</a>.

Bạn sẽ thấy tài liệu tương tác API được tạo tự động (cung cấp bởi <a href="https://github.com/swagger-api/swagger-ui" class="external-link" target="_blank">Swagger UI</a>):

#include "tensorflow/lite/schema/schema_generated.h"
#include "tensorflow/lite/toco/logging/conversion_log_util.h"
#include "tensorflow/lite/toco/logging/toco_conversion_log.pb.h"
#include "tensorflow/lite/toco/model.h"
#include "tensorflow/lite/toco/model_flags.pb.h"
#include "tensorflow/lite/toco/toco_convert.h"
#include "tensorflow/lite/toco/toco_flags.pb.h"
#include "tensorflow/lite/toco/toco_graphviz_dump_options.h"

#include "tensorflow/core/public/session.h"
#include "tensorflow/lite/toco/model_flags.pb.h"
#include "tensorflow/lite/toco/toco_flags.pb.h"
#include "tensorflow/lite/toco/types.pb.h"
#include "tsl/platform/statusor.h"

namespace tensorflow {
namespace internal {

// Register all custom ops including user specified custom ops.
Status RegisterAllCustomOps(const toco::TocoFlags& toco_flags);

      return DT_UINT32;
    case toco::IODataType::INT64:
      return DT_INT64;
    case toco::IODataType::UINT8:
      return DT_UINT8;
    case toco::IODataType::UINT64:
      return DT_UINT64;
    case toco::IODataType::STRING:
      return DT_STRING;
    case toco::IODataType::BOOL:
      return DT_BOOL;
    case toco::IODataType::COMPLEX64:
      return DT_COMPLEX64;

#include "tensorflow/core/platform/status.h"
#include "tensorflow/core/platform/types.h"
#include "tensorflow/lite/toco/model_flags.pb.h"
#include "tensorflow/lite/toco/toco_flags.pb.h"

namespace tensorflow {

// Converts the given Jax model to a TF Lite FlatBuffer
// string according to the given model flags, toco flags and tags. Returns error
// status if it fails to convert the input.

#include "tensorflow/lite/toco/model_flags.pb.h"
#include "tensorflow/lite/toco/toco_flags.pb.h"

namespace tensorflow {

// Converts the given saved_model(either v1 or v2) to a TF Lite FlatBuffer
// string according to the given model flags, toco flags and tags. Returns error
// status if it fails to convert the input.
Status ConvertSavedModelToTFLiteFlatBuffer(

    %0 = tfl.add %arg0, %arg1 {fused_activation_function = "RELU6", tac.device = "GPU"} : tensor<100xf32>
    func.return %0 : tensor<100xf32>
  }
  func.func @func_2_GPU_FLOAT(%arg0: tensor<100xf32>, %arg1: tensor<100xf32>) -> tensor<2x100xf32> attributes {tac.cost = 8.040000e+01 : f32, tac.device = "GPU", tac.inference_type = "FLOAT", tac.interface_name = "func_2"} {

// CHECK:           %[[VAL_3:.*]] = tfl.add %[[VAL_0]], %[[VAL_1]] {fused_activation_function = "RELU6", tac.device = "GPU", tac.inference_type = "FLOAT"} : tensor<1xf32>
// CHECK:           %[[VAL_4:.*]] = tfl.mul %[[VAL_3]], %[[VAL_2]] {fused_activation_function = "RELU6", tac.device = "GPU", tac.inference_type = "FLOAT"} : tensor<1xf32>