paddlepaddle--paddle/test/dygraph_to_static/test_ptb_lm.py

#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
#
# 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
#
#     http://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.

import logging
import time
import unittest

import numpy as np
from dygraph_to_static_utils import (
    Dy2StTestBase,
    enable_to_static_guard,
)

import paddle
from paddle.base.framework import unique_name
from paddle.optimizer import SGD

PRINT_STEP = 20
SEED = 2020


class SimpleLSTMRNN(paddle.nn.Layer):
    def __init__(
        self, hidden_size, num_steps, num_layers=2, init_scale=0.1, dropout=None
    ):
        super().__init__()
        self._hidden_size = hidden_size
        self._num_layers = num_layers
        self._init_scale = init_scale
        self._dropout = dropout
        self._num_steps = num_steps
        self.cell_array = []
        self.hidden_array = []

        self.weight_1_arr = []
        self.weight_2_arr = []
        self.bias_arr = []
        self.mask_array = []

        for i in range(self._num_layers):
            weight_1 = self.create_parameter(
                attr=paddle.ParamAttr(
                    initializer=paddle.nn.initializer.Uniform(
                        low=-self._init_scale, high=self._init_scale
                    )
                ),
                shape=[self._hidden_size * 2, self._hidden_size * 4],
                dtype="float32",
                default_initializer=paddle.nn.initializer.Uniform(
                    low=-self._init_scale, high=self._init_scale
                ),
            )
            self.weight_1_arr.append(self.add_parameter(f'w_{i}', weight_1))
            bias_1 = self.create_parameter(
                attr=paddle.ParamAttr(
                    initializer=paddle.nn.initializer.Uniform(
                        low=-self._init_scale, high=self._init_scale
                    )
                ),
                shape=[self._hidden_size * 4],
                dtype="float32",
                default_initializer=paddle.nn.initializer.Constant(0.0),
            )
            self.bias_arr.append(self.add_parameter(f'b_{i}', bias_1))

    def forward(self, input_embedding, init_hidden=None, init_cell=None):
        cell_array = []
        hidden_array = []

        for i in range(self._num_layers):
            hidden_array.append(init_hidden[i])
            cell_array.append(init_cell[i])

        res = []
        for index in range(self._num_steps):
            step_input = input_embedding[:, index, :]
            for k in range(self._num_layers):
                pre_hidden = hidden_array[k]
                pre_cell = cell_array[k]
                weight_1 = self.weight_1_arr[k]
                bias = self.bias_arr[k]

                nn = paddle.concat([step_input, pre_hidden], 1)
                gate_input = paddle.matmul(x=nn, y=weight_1)

                gate_input = paddle.add(gate_input, bias)
                i, j, f, o = paddle.split(
                    gate_input, num_or_sections=4, axis=-1
                )
                c = pre_cell * paddle.nn.functional.sigmoid(
                    f
                ) + paddle.nn.functional.sigmoid(i) * paddle.tanh(j)
                m = paddle.tanh(c) * paddle.nn.functional.sigmoid(o)
                hidden_array[k] = m
                cell_array[k] = c
                step_input = m

                if self._dropout is not None and self._dropout > 0.0:
                    step_input = paddle.nn.functional.dropout(
                        step_input,
                        p=self._dropout,
                        mode='upscale_in_train',
                    )
            res.append(step_input)
        real_res = paddle.concat(res, 1)
        real_res = paddle.reshape(
            real_res, [-1, self._num_steps, self._hidden_size]
        )
        last_hidden = paddle.concat(hidden_array, 1)
        last_hidden = paddle.reshape(
            last_hidden, shape=[-1, self._num_layers, self._hidden_size]
        )
        last_hidden = paddle.transpose(x=last_hidden, perm=[1, 0, 2])
        last_cell = paddle.concat(cell_array, 1)
        last_cell = paddle.reshape(
            last_cell, shape=[-1, self._num_layers, self._hidden_size]
        )
        last_cell = paddle.transpose(x=last_cell, perm=[1, 0, 2])
        return real_res, last_hidden, last_cell


class PtbModel(paddle.nn.Layer):
    def __init__(
        self,
        hidden_size,
        vocab_size,
        num_layers=2,
        num_steps=20,
        init_scale=0.1,
        dropout=None,
    ):
        super().__init__()
        self.hidden_size = hidden_size
        self.vocab_size = vocab_size
        self.init_scale = init_scale
        self.num_layers = num_layers
        self.num_steps = num_steps
        self.dropout = dropout
        self.simple_lstm_rnn = SimpleLSTMRNN(
            hidden_size,
            num_steps,
            num_layers=num_layers,
            init_scale=init_scale,
            dropout=dropout,
        )
        self.embedding = paddle.nn.Embedding(
            vocab_size,
            hidden_size,
            sparse=False,
            weight_attr=paddle.ParamAttr(
                name='embedding_para',
                initializer=paddle.nn.initializer.Uniform(
                    low=-init_scale, high=init_scale
                ),
            ),
        )
        self.softmax_weight = self.create_parameter(
            attr=paddle.ParamAttr(),
            shape=[self.hidden_size, self.vocab_size],
            dtype="float32",
            default_initializer=paddle.nn.initializer.Uniform(
                low=-self.init_scale, high=self.init_scale
            ),
        )
        self.softmax_bias = self.create_parameter(
            attr=paddle.ParamAttr(),
            shape=[self.vocab_size],
            dtype="float32",
            default_initializer=paddle.nn.initializer.Uniform(
                low=-self.init_scale, high=self.init_scale
            ),
        )

    def build_once(self, input, label, init_hidden, init_cell):
        pass

    def forward(self, input, label, init_hidden, init_cell):
        init_h = paddle.reshape(
            init_hidden, shape=[self.num_layers, -1, self.hidden_size]
        )

        init_c = paddle.reshape(
            init_cell, shape=[self.num_layers, -1, self.hidden_size]
        )

        x_emb = self.embedding(input)

        x_emb = paddle.reshape(
            x_emb, shape=[-1, self.num_steps, self.hidden_size]
        )
        if self.dropout is not None and self.dropout > 0.0:
            x_emb = paddle.nn.functional.dropout(
                x_emb,
                p=self.dropout,
                mode='upscale_in_train',
            )
        rnn_out, last_hidden, last_cell = self.simple_lstm_rnn(
            x_emb, init_h, init_c
        )

        projection = paddle.matmul(rnn_out, self.softmax_weight)
        projection = paddle.add(projection, self.softmax_bias)

        loss = paddle.nn.functional.cross_entropy(
            input=projection, label=label, soft_label=False, reduction="none"
        )
        loss = paddle.reshape(loss, shape=[-1, self.num_steps])
        loss = paddle.mean(loss, axis=[0])
        loss = paddle.sum(loss)

        return loss, last_hidden, last_cell

    def debug_emb(self):
        np.save("emb_grad", self.x_emb.gradient())


def train():
    num_layers = 1
    batch_size = 4
    hidden_size = 10
    num_steps = 3
    init_scale = 0.1
    max_epoch = 1
    dropout = 0.0
    vocab_size = 1000
    batch_num = 200

    with unique_name.guard():
        paddle.seed(SEED)
        paddle.framework.random._manual_program_seed(SEED)
        ptb_model = paddle.jit.to_static(
            PtbModel(
                hidden_size=hidden_size,
                vocab_size=vocab_size,
                num_layers=num_layers,
                num_steps=num_steps,
                init_scale=init_scale,
                dropout=dropout,
            )
        )

        sgd = SGD(learning_rate=1e-3, parameters=ptb_model.parameters())

        for epoch_id in range(max_epoch):
            total_loss = 0.0
            iters = 0.0
            total_sample = 0

            init_hidden_data = np.zeros(
                (num_layers, batch_size, hidden_size), dtype='float32'
            )
            init_cell_data = np.zeros(
                (num_layers, batch_size, hidden_size), dtype='float32'
            )

            init_hidden = paddle.to_tensor(init_hidden_data)
            init_cell = paddle.to_tensor(init_cell_data)
            for step_id in range(batch_num):
                x_data = np.arange(12).reshape(4, 3).astype('int64')
                y_data = np.arange(1, 13).reshape(4, 3).astype('int64')
                y_data = y_data.reshape((-1, 1))

                x_data = x_data.reshape((-1, num_steps, 1))
                y_data = y_data.reshape((-1, num_steps, 1))

                x = paddle.to_tensor(x_data)
                y = paddle.to_tensor(y_data)

                dy_loss, last_hidden, last_cell = ptb_model(
                    x, y, init_hidden, init_cell
                )
                out_loss = dy_loss.numpy()

                dy_loss.backward()
                sgd.minimize(dy_loss)
                ptb_model.clear_gradients()

                total_loss += out_loss
                iters += num_steps
                total_sample += 1
                if step_id % PRINT_STEP == 0:
                    if step_id == 0:
                        logging.info(
                            f"epoch {epoch_id} | step {step_id}, loss {total_loss / total_sample:0.3f}"
                        )
                        avg_batch_time = time.time()
                    else:
                        speed = PRINT_STEP / (time.time() - avg_batch_time)
                        logging.info(
                            f"epoch {epoch_id} | step {step_id}, loss {total_loss / total_sample:0.3f}, speed {speed:.3f} steps/s"
                        )
                        avg_batch_time = time.time()

        return out_loss, last_hidden.numpy(), last_cell.numpy()


def train_dygraph():
    with enable_to_static_guard(False):
        return train()


def train_static():
    return train()


class TestPtb(Dy2StTestBase):
    def test_check_result(self):
        loss_1, hidden_1, cell_1 = train_dygraph()
        loss_2, hidden_2, cell_2 = train_static()

        np.testing.assert_allclose(loss_1, loss_2, rtol=1e-05)
        np.testing.assert_allclose(hidden_1, hidden_2, rtol=1e-05)
        np.testing.assert_allclose(cell_1, cell_2, rtol=1e-05)


if __name__ == '__main__':
    unittest.main()