Colabを使った機械学習について

使用環境

Google Colab

実行時のコード
!python train.py --dataroot ./datasets/2200 --name 2200_batch_16 --label_nc 0 --no_instance --save_latest_freq 2200 --save_epoch_freq 100 --use_dropout --batchSize 16 --tf_log --lr 0.00005

やりたいこと

Google Colab上でPix2PixHDを用いて学習
tensorboardを使いたいためtensorflowを使用

発生している問題・エラーメッセージ

tensorboardを使いたいためtensorflowを用いてるのですが以下のようなエラーが出ており，対象方がわかりません．
AttributeError: module 'tensorflow' has no attribute 'Summary'

------------ Options -------------
batchSize: 16
beta1: 0.5
checkpoints_dir: ./checkpoints
continue_train: False
data_type: 32
dataroot: ./datasets/2200
debug: False
display_freq: 100
display_winsize: 512
feat_num: 3
fineSize: 512
fp16: False
gpu_ids: [0]
input_nc: 3
instance_feat: False
isTrain: True
label_feat: False
label_nc: 0
lambda_feat: 10.0
loadSize: 1024
load_features: False
load_pretrain: 
local_rank: 0
lr: 5e-05
max_dataset_size: inf
model: pix2pixHD
nThreads: 2
n_blocks_global: 9
n_blocks_local: 3
n_clusters: 10
n_downsample_E: 4
n_downsample_global: 4
n_layers_D: 3
n_local_enhancers: 1
name: 2200_batch_16
ndf: 64
nef: 16
netG: global
ngf: 64
niter: 100
niter_decay: 100
niter_fix_global: 0
no_flip: False
no_ganFeat_loss: False
no_html: False
no_instance: True
no_lsgan: False
no_vgg_loss: False
norm: instance
num_D: 2
output_nc: 3
phase: train
pool_size: 0
print_freq: 100
resize_or_crop: scale_width
save_epoch_freq: 100
save_latest_freq: 2200
serial_batches: False
tf_log: True
use_dropout: True
verbose: False
which_epoch: latest
-------------- End ----------------

  (scale1_layer4): Sequential(
    (0): Conv2d(512, 1, kernel_size=(4, 4), stride=(1, 1), padding=(2, 2))
  )
  (downsample): AvgPool2d(kernel_size=3, stride=2, padding=[1, 1])
)
/usr/local/lib/python3.7/dist-packages/torchvision/models/_utils.py:209: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and will be removed in 0.15, please use 'weights' instead.
  f"The parameter '{pretrained_param}' is deprecated since 0.13 and will be removed in 0.15, "
/usr/local/lib/python3.7/dist-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and will be removed in 0.15. The current behavior is equivalent to passing `weights=VGG19_Weights.IMAGENET1K_V1`. You can also use `weights=VGG19_Weights.DEFAULT` to get the most up-to-date weights.
  warnings.warn(msg)
2022-09-06 13:31:55.624112: W tensorflow/core/common_runtime/gpu/gpu_bfc_allocator.cc:39] Overriding allow_growth setting because the TF_FORCE_GPU_ALLOW_GROWTH environment variable is set. Original config value was 0.
create web directory ./checkpoints/2200_batch_16/web...
(epoch: 1, iters: 300, time: 0.555) G_GAN: 0.676 G_GAN_Feat: 3.975 G_VGG: 4.539 D_real: 0.624 D_fake: 0.616 
Traceback (most recent call last):
  File "train.py", line 104, in <module>
    visualizer.plot_current_errors(errors, total_steps)
  File "/content/drive/MyDrive/pix2pixHD/util/visualizer.py", line 99, in plot_current_errors
    summary = self.tf.Summary(value=[self.tf.Summary.Value(tag=tag, simple_value=value)])
AttributeError: module 'tensorflow' has no attribute 'Summary'

該当のソースコード

python
1import numpy as np
2import os
3import ntpath
4import time
5from . import util
6from . import html
7import scipy.misc
8try:
9    from StringIO import StringIO  # Python 2.7
10except ImportError:
11    from io import BytesIO         # Python 3.x
12
13class Visualizer():
14    def __init__(self, opt):
15        # self.opt = opt
16        self.tf_log = opt.tf_log
17        self.use_html = opt.isTrain and not opt.no_html
18        self.win_size = opt.display_winsize
19        self.name = opt.name
20        if self.tf_log:
21            import tensorflow as tf
22            self.tf = tf
23            self.log_dir = os.path.join(opt.checkpoints_dir, opt.name, 'logs')
24            # self.writer = tf.summary.FileWriter(self.log_dir)
25            self.writer = tf.summary.create_file_writer(self.log_dir)
26
27        if self.use_html:
28            self.web_dir = os.path.join(opt.checkpoints_dir, opt.name, 'web')
29            self.img_dir = os.path.join(self.web_dir, 'images')
30            print('create web directory %s...' % self.web_dir)
31            util.mkdirs([self.web_dir, self.img_dir])
32        self.log_name = os.path.join(opt.checkpoints_dir, opt.name, 'loss_log.txt')
33        with open(self.log_name, "a") as log_file:
34            now = time.strftime("%c")
35            log_file.write('================ Training Loss (%s) ================\n' % now)
36
37    # |visuals|: dictionary of images to display or save
38    def display_current_results(self, visuals, epoch, step):
39        if self.tf_log: # show images in tensorboard output
40            img_summaries = []
41            for label, image_numpy in visuals.items():
42                # Write the image to a string
43                try:
44                    s = StringIO()
45                except:
46                    s = BytesIO()
47                scipy.misc.toimage(image_numpy).save(s, format="jpeg")
48                # Create an Image object
49                img_sum = self.tf.Summary.Image(encoded_image_string=s.getvalue(), height=image_numpy.shape[0], width=image_numpy.shape[1])
50                # Create a Summary value
51                img_summaries.append(self.tf.Summary.Value(tag=label, image=img_sum))
52
53            # Create and write Summary
54            summary = self.tf.Summary(value=img_summaries)
55            self.writer.add_summary(summary, step)
56
57        if self.use_html: # save images to a html file
58            for label, image_numpy in visuals.items():
59                if isinstance(image_numpy, list):
60                    for i in range(len(image_numpy)):
61                        img_path = os.path.join(self.img_dir, 'epoch%.3d_%s_%d.jpg' % (epoch, label, i))
62                        util.save_image(image_numpy[i], img_path)
63                else:
64                    img_path = os.path.join(self.img_dir, 'epoch%.3d_%s.jpg' % (epoch, label))
65                    util.save_image(image_numpy, img_path)
66
67            # update website
68            webpage = html.HTML(self.web_dir, 'Experiment name = %s' % self.name, refresh=30)
69            for n in range(epoch, 0, -1):
70                webpage.add_header('epoch [%d]' % n)
71                ims = []
72                txts = []
73                links = []
74
75                for label, image_numpy in visuals.items():
76                    if isinstance(image_numpy, list):
77                        for i in range(len(image_numpy)):
78                            img_path = 'epoch%.3d_%s_%d.jpg' % (n, label, i)
79                            ims.append(img_path)
80                            txts.append(label+str(i))
81                            links.append(img_path)
82                    else:
83                        img_path = 'epoch%.3d_%s.jpg' % (n, label)
84                        ims.append(img_path)
85                        txts.append(label)
86                        links.append(img_path)
87                if len(ims) < 10:
88                    webpage.add_images(ims, txts, links, width=self.win_size)
89                else:
90                    num = int(round(len(ims)/2.0))
91                    webpage.add_images(ims[:num], txts[:num], links[:num], width=self.win_size)
92                    webpage.add_images(ims[num:], txts[num:], links[num:], width=self.win_size)
93            webpage.save()
94
95    # errors: dictionary of error labels and values
96    def plot_current_errors(self, errors, step):
97        if self.tf_log:
98            for tag, value in errors.items():
99                # summary = self.tf.Summary(value=[self.tf.Summary.Value(tag=tag, simple_value=value)])
100                summary = self.tf.summary.create_file_writer(value=[self.tf.compat.v2.Summary.Value(tag=tag, simple_value=value)])
101               
102                self.writer.add_summary(summary, step)
103
104    # errors: same format as |errors| of plotCurrentErrors
105    def print_current_errors(self, epoch, i, errors, t):
106        message = '(epoch: %d, iters: %d, time: %.3f) ' % (epoch, i, t)
107        for k, v in errors.items():
108            if v != 0:
109                message += '%s: %.3f ' % (k, v)
110
111        print(message)
112        with open(self.log_name, "a") as log_file:
113            log_file.write('%s\n' % message)
114
115    # save image to the disk
116    def save_images(self, webpage, visuals, image_path):
117        image_dir = webpage.get_image_dir()
118        short_path = ntpath.basename(image_path[0])
119        name = os.path.splitext(short_path)[0]
120
121        webpage.add_header(name)
122        ims = []
123        txts = []
124        links = []
125
126        for label, image_numpy in visuals.items():
127            image_name = '%s_%s.jpg' % (name, label)
128            save_path = os.path.join(image_dir, image_name)
129            util.save_image(image_numpy, save_path)
130
131            ims.append(image_name)
132            txts.append(label)
133            links.append(image_name)
134        webpage.add_images(ims, txts, links, width=self.win_size)
135

試したこと

tensorflowのバージョン違いによる書き換えが必要という記事を読み修正したのですが，変更後もエラーが出ます

補足

tensorboard 2.8.0
tensorflow 2.8.2

です

jbpb0

2022/09/07 05:34 編集

質問に記載のコードは https://github.com/NVIDIA/pix2pixHD/blob/master/util/visualizer.py のようですが、これの日付はLatest commit 5a2c872 on 14 Jun 2019で、tensorflow 2.0 https://pypi.org/project/tensorflow/2.0.0/ のリリース日のReleased: Oct 1, 2019よりも前なので、tensorflow 1.*用に書かれたものだと思うので、google colabにtensorflow 1.*を入れたら、この質問のエラーは出なくなるのではないですかね https://prglog.info/home/?p=22