program.py 20.3 KB
Newer Older
M
refine  
MissPenguin 已提交
1
# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
L
LDOUBLEV 已提交
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
#
# 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.

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

W
WenmuZhou 已提交
19
import os
L
LDOUBLEV 已提交
20
import sys
21
import platform
L
LDOUBLEV 已提交
22 23
import yaml
import time
24
import datetime
W
WenmuZhou 已提交
25 26 27 28 29
import paddle
import paddle.distributed as dist
from tqdm import tqdm
from argparse import ArgumentParser, RawDescriptionHelpFormatter

L
LDOUBLEV 已提交
30 31
from ppocr.utils.stats import TrainingStats
from ppocr.utils.save_load import save_model
32
from ppocr.utils.utility import print_dict, AverageMeter
D
dyning 已提交
33
from ppocr.utils.logging import get_logger
L
LDOUBLEV 已提交
34
from ppocr.utils import profiler
D
dyning 已提交
35
from ppocr.data import build_dataloader
L
LDOUBLEV 已提交
36

D
dyning 已提交
37

L
LDOUBLEV 已提交
38 39 40 41 42 43 44
class ArgsParser(ArgumentParser):
    def __init__(self):
        super(ArgsParser, self).__init__(
            formatter_class=RawDescriptionHelpFormatter)
        self.add_argument("-c", "--config", help="configuration file to use")
        self.add_argument(
            "-o", "--opt", nargs='+', help="set configuration options")
L
LDOUBLEV 已提交
45 46 47 48 49
        self.add_argument(
            '-p',
            '--profiler_options',
            type=str,
            default=None,
50 51
            help='The option of profiler, which should be in format ' \
                 '\"key1=value1;key2=value2;key3=value3\".'
L
LDOUBLEV 已提交
52
        )
L
LDOUBLEV 已提交
53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80

    def parse_args(self, argv=None):
        args = super(ArgsParser, self).parse_args(argv)
        assert args.config is not None, \
            "Please specify --config=configure_file_path."
        args.opt = self._parse_opt(args.opt)
        return args

    def _parse_opt(self, opts):
        config = {}
        if not opts:
            return config
        for s in opts:
            s = s.strip()
            k, v = s.split('=')
            config[k] = yaml.load(v, Loader=yaml.Loader)
        return config


def load_config(file_path):
    """
    Load config from yml/yaml file.
    Args:
        file_path (str): Path of the config file to be loaded.
    Returns: global config
    """
    _, ext = os.path.splitext(file_path)
    assert ext in ['.yml', '.yaml'], "only support yaml files for now"
81 82
    config = yaml.load(open(file_path, 'rb'), Loader=yaml.Loader)
    return config
L
LDOUBLEV 已提交
83 84


85
def merge_config(config, opts):
L
LDOUBLEV 已提交
86 87 88 89 90 91
    """
    Merge config into global config.
    Args:
        config (dict): Config to be merged.
    Returns: global config
    """
92
    for key, value in opts.items():
L
LDOUBLEV 已提交
93
        if "." not in key:
94 95
            if isinstance(value, dict) and key in config:
                config[key].update(value)
L
LDOUBLEV 已提交
96
            else:
97
                config[key] = value
L
LDOUBLEV 已提交
98 99
        else:
            sub_keys = key.split('.')
T
tink2123 已提交
100
            assert (
101
                sub_keys[0] in config
102 103
            ), "the sub_keys can only be one of global_config: {}, but get: " \
               "{}, please check your running command".format(
104 105
                config.keys(), sub_keys[0])
            cur = config[sub_keys[0]]
L
LDOUBLEV 已提交
106 107 108 109 110
            for idx, sub_key in enumerate(sub_keys[1:]):
                if idx == len(sub_keys) - 2:
                    cur[sub_key] = value
                else:
                    cur = cur[sub_key]
111
    return config
L
LDOUBLEV 已提交
112 113 114 115 116 117 118 119 120 121 122 123 124 125


def check_gpu(use_gpu):
    """
    Log error and exit when set use_gpu=true in paddlepaddle
    cpu version.
    """
    err = "Config use_gpu cannot be set as true while you are " \
          "using paddlepaddle cpu version ! \nPlease try: \n" \
          "\t1. Install paddlepaddle-gpu to run model on GPU \n" \
          "\t2. Set use_gpu as false in config file to run " \
          "model on CPU"

    try:
W
WenmuZhou 已提交
126
        if use_gpu and not paddle.is_compiled_with_cuda():
W
WenmuZhou 已提交
127
            print(err)
L
LDOUBLEV 已提交
128 129 130 131 132
            sys.exit(1)
    except Exception as e:
        pass


133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151
def check_xpu(use_xpu):
    """
    Log error and exit when set use_xpu=true in paddlepaddle
    cpu/gpu version.
    """
    err = "Config use_xpu cannot be set as true while you are " \
          "using paddlepaddle cpu/gpu version ! \nPlease try: \n" \
          "\t1. Install paddlepaddle-xpu to run model on XPU \n" \
          "\t2. Set use_xpu as false in config file to run " \
          "model on CPU/GPU"

    try:
        if use_xpu and not paddle.is_compiled_with_xpu():
            print(err)
            sys.exit(1)
    except Exception as e:
        pass


W
WenmuZhou 已提交
152
def train(config,
D
dyning 已提交
153 154 155
          train_dataloader,
          valid_dataloader,
          device,
W
WenmuZhou 已提交
156 157 158 159 160 161 162 163
          model,
          loss_class,
          optimizer,
          lr_scheduler,
          post_process_class,
          eval_class,
          pre_best_model_dict,
          logger,
S
stephon 已提交
164 165
          vdl_writer=None,
          scaler=None):
W
WenmuZhou 已提交
166 167
    cal_metric_during_train = config['Global'].get('cal_metric_during_train',
                                                   False)
168
    calc_epoch_interval = config['Global'].get('calc_epoch_interval', 1)
L
LDOUBLEV 已提交
169 170 171 172
    log_smooth_window = config['Global']['log_smooth_window']
    epoch_num = config['Global']['epoch_num']
    print_batch_step = config['Global']['print_batch_step']
    eval_batch_step = config['Global']['eval_batch_step']
L
LDOUBLEV 已提交
173
    profiler_options = config['profiler_options']
W
WenmuZhou 已提交
174

D
dyning 已提交
175
    global_step = 0
176 177
    if 'global_step' in pre_best_model_dict:
        global_step = pre_best_model_dict['global_step']
L
LDOUBLEV 已提交
178 179 180 181
    start_eval_step = 0
    if type(eval_batch_step) == list and len(eval_batch_step) >= 2:
        start_eval_step = eval_batch_step[0]
        eval_batch_step = eval_batch_step[1]
W
WenmuZhou 已提交
182 183
        if len(valid_dataloader) == 0:
            logger.info(
184 185
                'No Images in eval dataset, evaluation during training ' \
                'will be disabled'
W
WenmuZhou 已提交
186 187
            )
            start_eval_step = 1e111
L
LDOUBLEV 已提交
188
        logger.info(
189 190
            "During the training process, after the {}th iteration, " \
            "an evaluation is run every {} iterations".
L
LDOUBLEV 已提交
191
            format(start_eval_step, eval_batch_step))
L
LDOUBLEV 已提交
192 193
    save_epoch_step = config['Global']['save_epoch_step']
    save_model_dir = config['Global']['save_model_dir']
194 195
    if not os.path.exists(save_model_dir):
        os.makedirs(save_model_dir)
W
WenmuZhou 已提交
196 197 198 199
    main_indicator = eval_class.main_indicator
    best_model_dict = {main_indicator: 0}
    best_model_dict.update(pre_best_model_dict)
    train_stats = TrainingStats(log_smooth_window, ['lr'])
T
tink2123 已提交
200
    model_average = False
W
WenmuZhou 已提交
201 202
    model.train()

T
tink2123 已提交
203
    use_srn = config['Architecture']['algorithm'] == "SRN"
T
tink2123 已提交
204
    extra_input = config['Architecture'][
L
LDOUBLEV 已提交
205
        'algorithm'] in ["SRN", "NRTR", "SAR", "SEED"]
206
    try:
L
fix bug  
LDOUBLEV 已提交
207
        model_type = config['Architecture']['model_type']
208
    except:
L
fix bug  
LDOUBLEV 已提交
209
        model_type = None
T
tink2123 已提交
210
    algorithm = config['Architecture']['algorithm']
T
tink2123 已提交
211

212 213 214 215
    start_epoch = best_model_dict[
        'start_epoch'] if 'start_epoch' in best_model_dict else 1

    total_samples = 0
216 217
    train_reader_cost = 0.0
    train_batch_cost = 0.0
218
    reader_start = time.time()
219
    eta_meter = AverageMeter()
220 221 222

    max_iter = len(train_dataloader) - 1 if platform.system(
    ) == "Windows" else len(train_dataloader)
W
WenmuZhou 已提交
223

T
tink2123 已提交
224
    for epoch in range(start_epoch, epoch_num + 1):
225 226 227 228 229
        if train_dataloader.dataset.need_reset:
            train_dataloader = build_dataloader(
                config, 'Train', device, logger, seed=epoch)
            max_iter = len(train_dataloader) - 1 if platform.system(
            ) == "Windows" else len(train_dataloader)
W
WenmuZhou 已提交
230
        for idx, batch in enumerate(train_dataloader):
L
LDOUBLEV 已提交
231
            profiler.add_profiler_step(profiler_options)
文幕地方's avatar
文幕地方 已提交
232
            train_reader_cost += time.time() - reader_start
J
Jane-Ding 已提交
233
            if idx >= max_iter:
W
WenmuZhou 已提交
234 235 236
                break
            lr = optimizer.get_lr()
            images = batch[0]
T
tink2123 已提交
237
            if use_srn:
T
tink2123 已提交
238
                model_average = True
S
stephon 已提交
239 240 241 242 243 244 245 246

            # use amp
            if scaler:
                with paddle.amp.auto_cast():
                    if model_type == 'table' or extra_input:
                        preds = model(images, data=batch[1:])
                    else:
                        preds = model(images)
T
tink2123 已提交
247
            else:
S
stephon 已提交
248 249
                if model_type == 'table' or extra_input:
                    preds = model(images, data=batch[1:])
250
                elif model_type in ["kie", 'vqa']:
L
LDOUBLEV 已提交
251
                    preds = model(batch)
S
stephon 已提交
252 253
                else:
                    preds = model(images)
254

W
WenmuZhou 已提交
255 256
            loss = loss_class(preds, batch)
            avg_loss = loss['loss']
S
stephon 已提交
257 258 259 260 261 262 263 264

            if scaler:
                scaled_avg_loss = scaler.scale(avg_loss)
                scaled_avg_loss.backward()
                scaler.minimize(optimizer, scaled_avg_loss)
            else:
                avg_loss.backward()
                optimizer.step()
W
WenmuZhou 已提交
265
            optimizer.clear_grad()
W
WenmuZhou 已提交
266

267 268 269 270 271 272 273 274 275 276
            if cal_metric_during_train and epoch % calc_epoch_interval == 0:  # only rec and cls need
                batch = [item.numpy() for item in batch]
                if model_type in ['table', 'kie']:
                    eval_class(preds, batch)
                else:
                    post_result = post_process_class(preds, batch[1])
                    eval_class(post_result, batch)
                metric = eval_class.get_metric()
                train_stats.update(metric)

277 278 279
            train_batch_time = time.time() - reader_start
            train_batch_cost += train_batch_time
            eta_meter.update(train_batch_time)
280
            global_step += 1
文幕地方's avatar
文幕地方 已提交
281
            total_samples += len(images)
W
WenmuZhou 已提交
282

D
dyning 已提交
283 284
            if not isinstance(lr_scheduler, float):
                lr_scheduler.step()
W
WenmuZhou 已提交
285 286 287 288 289 290 291 292 293 294 295

            # logger and visualdl
            stats = {k: v.numpy().mean() for k, v in loss.items()}
            stats['lr'] = lr
            train_stats.update(stats)

            if vdl_writer is not None and dist.get_rank() == 0:
                for k, v in train_stats.get().items():
                    vdl_writer.add_scalar('TRAIN/{}'.format(k), v, global_step)
                vdl_writer.add_scalar('TRAIN/lr', lr, global_step)

296 297 298
            if dist.get_rank() == 0 and (
                (global_step > 0 and global_step % print_batch_step == 0) or
                (idx >= len(train_dataloader) - 1)):
W
WenmuZhou 已提交
299
                logs = train_stats.log()
L
LDOUBLEV 已提交
300

301 302 303 304 305
                eta_sec = ((epoch_num + 1 - epoch) * \
                    len(train_dataloader) - idx - 1) * eta_meter.avg
                eta_sec_format = str(datetime.timedelta(seconds=int(eta_sec)))
                strs = 'epoch: [{}/{}], global_step: {}, {}, avg_reader_cost: ' \
                       '{:.5f} s, avg_batch_cost: {:.5f} s, avg_samples: {}, ' \
L
LDOUBLEV 已提交
306
                       'ips: {:.5f} samples/s, eta: {}'.format(
307 308 309 310 311
                    epoch, epoch_num, global_step, logs,
                    train_reader_cost / print_batch_step,
                    train_batch_cost / print_batch_step,
                    total_samples / print_batch_step,
                    total_samples / train_batch_cost, eta_sec_format)
W
WenmuZhou 已提交
312
                logger.info(strs)
313

文幕地方's avatar
文幕地方 已提交
314
                total_samples = 0
315 316
                train_reader_cost = 0.0
                train_batch_cost = 0.0
W
WenmuZhou 已提交
317 318
            # eval
            if global_step > start_eval_step and \
319 320
                    (global_step - start_eval_step) % eval_batch_step == 0 \
                    and dist.get_rank() == 0:
T
tink2123 已提交
321 322 323 324 325 326 327
                if model_average:
                    Model_Average = paddle.incubate.optimizer.ModelAverage(
                        0.15,
                        parameters=model.parameters(),
                        min_average_window=10000,
                        max_average_window=15625)
                    Model_Average.apply()
T
tink2123 已提交
328 329 330 331 332
                cur_metric = eval(
                    model,
                    valid_dataloader,
                    post_process_class,
                    eval_class,
M
refine  
MissPenguin 已提交
333
                    model_type,
T
tink2123 已提交
334
                    extra_input=extra_input)
L
LDOUBLEV 已提交
335 336 337
                cur_metric_str = 'cur metric, {}'.format(', '.join(
                    ['{}: {}'.format(k, v) for k, v in cur_metric.items()]))
                logger.info(cur_metric_str)
W
WenmuZhou 已提交
338 339 340

                # logger metric
                if vdl_writer is not None:
L
LDOUBLEV 已提交
341
                    for k, v in cur_metric.items():
W
WenmuZhou 已提交
342 343
                        if isinstance(v, (float, int)):
                            vdl_writer.add_scalar('EVAL/{}'.format(k),
L
LDOUBLEV 已提交
344 345
                                                  cur_metric[k], global_step)
                if cur_metric[main_indicator] >= best_model_dict[
W
WenmuZhou 已提交
346
                        main_indicator]:
L
LDOUBLEV 已提交
347
                    best_model_dict.update(cur_metric)
W
WenmuZhou 已提交
348 349 350 351 352 353
                    best_model_dict['best_epoch'] = epoch
                    save_model(
                        model,
                        optimizer,
                        save_model_dir,
                        logger,
354
                        config,
W
WenmuZhou 已提交
355 356 357
                        is_best=True,
                        prefix='best_accuracy',
                        best_model_dict=best_model_dict,
358 359
                        epoch=epoch,
                        global_step=global_step)
L
LDOUBLEV 已提交
360
                best_str = 'best metric, {}'.format(', '.join([
W
WenmuZhou 已提交
361 362 363 364 365 366 367 368
                    '{}: {}'.format(k, v) for k, v in best_model_dict.items()
                ]))
                logger.info(best_str)
                # logger best metric
                if vdl_writer is not None:
                    vdl_writer.add_scalar('EVAL/best_{}'.format(main_indicator),
                                          best_model_dict[main_indicator],
                                          global_step)
369

文幕地方's avatar
文幕地方 已提交
370
            reader_start = time.time()
W
WenmuZhou 已提交
371 372 373 374 375 376
        if dist.get_rank() == 0:
            save_model(
                model,
                optimizer,
                save_model_dir,
                logger,
377
                config,
W
WenmuZhou 已提交
378 379 380
                is_best=False,
                prefix='latest',
                best_model_dict=best_model_dict,
381 382
                epoch=epoch,
                global_step=global_step)
W
WenmuZhou 已提交
383 384 385 386 387 388
        if dist.get_rank() == 0 and epoch > 0 and epoch % save_epoch_step == 0:
            save_model(
                model,
                optimizer,
                save_model_dir,
                logger,
389
                config,
W
WenmuZhou 已提交
390 391 392
                is_best=False,
                prefix='iter_epoch_{}'.format(epoch),
                best_model_dict=best_model_dict,
393 394
                epoch=epoch,
                global_step=global_step)
L
LDOUBLEV 已提交
395
    best_str = 'best metric, {}'.format(', '.join(
W
WenmuZhou 已提交
396 397 398 399
        ['{}: {}'.format(k, v) for k, v in best_model_dict.items()]))
    logger.info(best_str)
    if dist.get_rank() == 0 and vdl_writer is not None:
        vdl_writer.close()
L
LDOUBLEV 已提交
400 401 402
    return


M
refine  
MissPenguin 已提交
403 404 405 406
def eval(model,
         valid_dataloader,
         post_process_class,
         eval_class,
L
LDOUBLEV 已提交
407
         model_type=None,
T
tink2123 已提交
408
         extra_input=False):
W
WenmuZhou 已提交
409 410 411 412
    model.eval()
    with paddle.no_grad():
        total_frame = 0.0
        total_time = 0.0
文幕地方's avatar
文幕地方 已提交
413 414 415 416 417
        pbar = tqdm(
            total=len(valid_dataloader),
            desc='eval model:',
            position=0,
            leave=True)
418 419
        max_iter = len(valid_dataloader) - 1 if platform.system(
        ) == "Windows" else len(valid_dataloader)
W
WenmuZhou 已提交
420
        for idx, batch in enumerate(valid_dataloader):
421
            if idx >= max_iter:
W
WenmuZhou 已提交
422
                break
W
fix bug  
WenmuZhou 已提交
423
            images = batch[0]
W
WenmuZhou 已提交
424
            start = time.time()
T
tink2123 已提交
425
            if model_type == 'table' or extra_input:
M
refine  
MissPenguin 已提交
426
                preds = model(images, data=batch[1:])
427
            elif model_type in ["kie", 'vqa']:
L
LDOUBLEV 已提交
428
                preds = model(batch)
X
xiaoting 已提交
429
            else:
L
LDOUBLEV 已提交
430
                preds = model(images)
431 432 433 434 435 436 437

            batch_numpy = []
            for item in batch:
                if isinstance(item, paddle.Tensor):
                    batch_numpy.append(item.numpy())
                else:
                    batch_numpy.append(item)
W
WenmuZhou 已提交
438 439 440
            # Obtain usable results from post-processing methods
            total_time += time.time() - start
            # Evaluate the results of the current batch
L
LDOUBLEV 已提交
441
            if model_type in ['table', 'kie']:
442 443 444 445
                eval_class(preds, batch_numpy)
            elif model_type in ['vqa']:
                post_result = post_process_class(preds, batch_numpy)
                eval_class(post_result, batch_numpy)
M
MissPenguin 已提交
446
            else:
447 448
                post_result = post_process_class(preds, batch_numpy[1])
                eval_class(post_result, batch_numpy)
L
LDOUBLEV 已提交
449

W
fix bug  
WenmuZhou 已提交
450
            pbar.update(1)
W
WenmuZhou 已提交
451
            total_frame += len(images)
L
LDOUBLEV 已提交
452 453
        # Get final metric,eg. acc or hmean
        metric = eval_class.get_metric()
D
dyning 已提交
454

W
fix bug  
WenmuZhou 已提交
455
    pbar.close()
W
WenmuZhou 已提交
456
    model.train()
L
LDOUBLEV 已提交
457 458
    metric['fps'] = total_frame / total_time
    return metric
L
licx 已提交
459

T
tink2123 已提交
460

B
Bin Lu 已提交
461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509
def update_center(char_center, post_result, preds):
    result, label = post_result
    feats, logits = preds
    logits = paddle.argmax(logits, axis=-1)
    feats = feats.numpy()
    logits = logits.numpy()

    for idx_sample in range(len(label)):
        if result[idx_sample][0] == label[idx_sample][0]:
            feat = feats[idx_sample]
            logit = logits[idx_sample]
            for idx_time in range(len(logit)):
                index = logit[idx_time]
                if index in char_center.keys():
                    char_center[index][0] = (
                        char_center[index][0] * char_center[index][1] +
                        feat[idx_time]) / (char_center[index][1] + 1)
                    char_center[index][1] += 1
                else:
                    char_center[index] = [feat[idx_time], 1]
    return char_center


def get_center(model, eval_dataloader, post_process_class):
    pbar = tqdm(total=len(eval_dataloader), desc='get center:')
    max_iter = len(eval_dataloader) - 1 if platform.system(
    ) == "Windows" else len(eval_dataloader)
    char_center = dict()
    for idx, batch in enumerate(eval_dataloader):
        if idx >= max_iter:
            break
        images = batch[0]
        start = time.time()
        preds = model(images)

        batch = [item.numpy() for item in batch]
        # Obtain usable results from post-processing methods
        post_result = post_process_class(preds, batch[1])

        #update char_center
        char_center = update_center(char_center, post_result, preds)
        pbar.update(1)

    pbar.close()
    for key in char_center.keys():
        char_center[key] = char_center[key][0]
    return char_center


510
def preprocess(is_train=False):
L
licx 已提交
511
    FLAGS = ArgsParser().parse_args()
L
LDOUBLEV 已提交
512
    profiler_options = FLAGS.profiler_options
L
licx 已提交
513
    config = load_config(FLAGS.config)
514
    config = merge_config(config, FLAGS.opt)
L
LDOUBLEV 已提交
515
    profile_dic = {"profiler_options": FLAGS.profiler_options}
516
    config = merge_config(config, profile_dic)
L
licx 已提交
517

W
WenmuZhou 已提交
518 519 520 521 522 523 524 525 526 527 528
    if is_train:
        # save_config
        save_model_dir = config['Global']['save_model_dir']
        os.makedirs(save_model_dir, exist_ok=True)
        with open(os.path.join(save_model_dir, 'config.yml'), 'w') as f:
            yaml.dump(
                dict(config), f, default_flow_style=False, sort_keys=False)
        log_file = '{}/train.log'.format(save_model_dir)
    else:
        log_file = None
    logger = get_logger(name='root', log_file=log_file)
L
licx 已提交
529 530 531 532 533

    # check if set use_gpu=True in paddlepaddle cpu version
    use_gpu = config['Global']['use_gpu']
    check_gpu(use_gpu)

534 535 536 537 538 539
    # check if set use_xpu=True in paddlepaddle cpu/gpu version
    use_xpu = False
    if 'use_xpu' in config['Global']:
        use_xpu = config['Global']['use_xpu']
    check_xpu(use_xpu)

W
WenmuZhou 已提交
540 541
    alg = config['Architecture']['algorithm']
    assert alg in [
J
Jethong 已提交
542
        'EAST', 'DB', 'SAST', 'Rosetta', 'CRNN', 'STARNet', 'RARE', 'SRN',
T
tink2123 已提交
543
        'CLS', 'PGNet', 'Distillation', 'NRTR', 'TableAttn', 'SAR', 'PSE',
544
        'SEED', 'SDMGR', 'LayoutXLM', 'LayoutLM', 'PREN', 'FCE'
W
WenmuZhou 已提交
545
    ]
L
licx 已提交
546

547 548 549 550 551
    device = 'cpu'
    if use_gpu:
        device = 'gpu:{}'.format(dist.ParallelEnv().dev_id)
    if use_xpu:
        device = 'xpu'
W
WenmuZhou 已提交
552
    device = paddle.set_device(device)
D
dyning 已提交
553

D
dyning 已提交
554
    config['Global']['distributed'] = dist.get_world_size() != 1
W
WenmuZhou 已提交
555

littletomatodonkey's avatar
littletomatodonkey 已提交
556
    if config['Global']['use_visualdl'] and dist.get_rank() == 0:
D
dyning 已提交
557
        from visualdl import LogWriter
L
fix bug  
LDOUBLEV 已提交
558
        save_model_dir = config['Global']['save_model_dir']
D
dyning 已提交
559 560 561 562 563 564 565 566 567
        vdl_writer_path = '{}/vdl/'.format(save_model_dir)
        os.makedirs(vdl_writer_path, exist_ok=True)
        vdl_writer = LogWriter(logdir=vdl_writer_path)
    else:
        vdl_writer = None
    print_dict(config, logger)
    logger.info('train with paddle {} and device {}'.format(paddle.__version__,
                                                            device))
    return config, device, logger, vdl_writer