transform.h¶

Namespaces¶

Name
tenzr
tenzr::optim

Classes¶

	Name
struct	tenzr::optim::OptState Optimizer state holding the step counter and per-parameter tensors.
class	tenzr::optim::GradientTransformation Abstract gradient transformation defining Init and Update for an optimizer.

Source code¶

// Copyright 2026 The Tenzr Project Authors
//
// SPDX-License-Identifier: Apache-2.0
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//      https://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

#ifndef TENZR_MODULES_OPTIM_TRANSFORM_H_
#define TENZR_MODULES_OPTIM_TRANSFORM_H_

#include <cstdint>
#include <functional>
#include <memory>
#include <string>
#include <unordered_map>
#include <vector>

#include "tenzr/frontend/tensor.h"

namespace tenzr::optim {

  struct OptState {
    uint64_t step{0};
    std::unordered_map<std::string, std::unordered_map<std::string, Tensor>> param_states;
  };

  class GradientTransformation {
   public:
    virtual ~GradientTransformation() = default;

    virtual OptState Init(const std::unordered_map<std::string, Tensor>& params) = 0;

    virtual std::unordered_map<std::string, Tensor> Update(const std::unordered_map<std::string, Tensor>& grads, OptState& state,
                                                           const std::unordered_map<std::string, Tensor>& params) = 0;
  };

}  // namespace tenzr::optim

#endif  // TENZR_MODULES_OPTIM_TRANSFORM_H_