diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/.keep b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/.keep new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/Basic_Model.py b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/Basic_Model.py new file mode 100644 index 0000000000000000000000000000000000000000..700d4e99d9e43066e5531187ab5664a026686a2f --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/Basic_Model.py @@ -0,0 +1,170 @@ +# Copyright 2017 The TensorFlow 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. +# ============================================================================ +# Copyright 2021 Huawei Technologies Co., Ltd +# +# 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 npu_bridge.npu_init import * +import tensorflow as tf +from tensorflow.contrib import layers +import sys + +class basic_network(object): + def __init__(self, cfg): + self.training=True + self.cfg = cfg + self.params_count = 0 #the amount of parameters + def init_params(self, *args, **kwargs): + def _variable_on_cpu(w_shape, b_shape, weight_decay=0.99, use_bias=True, name="conv"): + with tf.device('/cpu:0'): + w = tf.Variable(tf.truncated_normal(w_shape, 0.0, 0.001), trainable=True, name="%s_w" % name) + tf.add_to_collection(name="weights_l2_loss", value=self.calc_l1_loss(w, weight_decay)) + b = tf.Variable(tf.zeros(b_shape), trainable=use_bias, name="%s_b" % name) + return w, b + kernel_size = kwargs["kernel_size"] + in_channels = kwargs["in_channels"] + out_channels = kwargs["out_channels"] + # weight_decay = kwargs["weight_decay"] + w_shape = [kernel_size, kernel_size, in_channels, out_channels] + b_shape = [out_channels] + name = kwargs["name"] + self.params_count += kernel_size*kernel_size*in_channels*out_channels + self.params_count += out_channels + return _variable_on_cpu(w_shape, b_shape, use_bias=kwargs["use_bias"], name=name) + + def calc_loss(self, *args, **kwargs): + loss_type = kwargs["loss_type"] + x = kwargs["x"] + y = kwargs["y"] + if loss_type == "L1": + return tf.reduce_sum(tf.abs(x-y), name="L1_loss") + elif loss_type == "L2": + return tf.nn.l2_loss((x-y), name="L2_loss") + + def activation(self, *args, **kwargs): + act_type = kwargs["act_type"] + act_type = act_type.lower() + if act_type == "relu": + return tf.nn.relu(args[0]) + elif act_type == "lrelu": + slope = kwargs["slope"] + y = slope*args[0] + return tf.maximum(args[0], y) + elif act_type == "prelu": + return tf.nn.leaky_relu(args[0], alpha=0.2) + elif act_type == "tanh": + return tf.nn.tanh(args[0]) + else: + return args[0] + + def calc_l2_loss(self, weight, weight_decay): + _, _, _, outchannel = weight.get_shape().as_list() + return (weight_decay) * tf.reduce_sum(tf.square(weight)) / outchannel + + def calc_l1_loss(self, weight, weight_decay): + _, _, _, outchannel = weight.get_shape().as_list() + return (weight_decay)*tf.reduce_sum(tf.abs(weight)) / outchannel + + def batch_norm(self, *args, **kwargs): + return tf.layers.batch_normalization(args[0], training=kwargs["training"]) + + def instance_norm(self, *args, **kwargs): + return layers.instance_norm(args[0], kwargs["name"]) + + def hard_sigmoid(self, x): + return tf.nn.relu6((x+3)/6) + + def hard_swish(self, x): + return x * self.hard_sigmoid(x) + + def global_average_pooling(self, x, name="GAP"): + return tf.reduce_mean(x, axis=[1, 2], keep_dims=True, name="Global_Average_Pooling_%s" % name)#不降维 + + + def ConvBlock(self,x, in_channels, out_channels, kernel_size, stride=1, name="ConvBlock", + BN=True, use_bias=True, padding="VALID", act_type="relu", mode="CNA"): + + + assert (mode in ['CNA', 'NAC']), '[ERROR] Wrong mode in [%s]!' % sys.modules[__name__]#断言 + weight, bias = self.init_params(kernel_size=kernel_size, in_channels=in_channels, + out_channels=out_channels, use_bias=use_bias, name=name) + if mode == "CNA": + x = tf.nn.conv2d(x, filter=weight, strides=[1, stride, stride, 1], padding=padding) + x = tf.nn.bias_add(x, bias) + if BN: + if self.cfg.BN_type == "BN": + x = self.batch_norm(x, training=self.cfg.istrain) + elif self.cfg.BN_type == "IN": + x = self.instance_norm(x, name="%s_IN"%name) + else: + raise NotImplementedError('[ERROR] BN type [%s] is not implemented!' % self.cfg.BN_type) + x = self.activation(x, act_type=act_type) + return x + elif mode=="NAC": + if BN: + if self.cfg.BN_type == "BN": + x = self.batch_norm(x, training=self.cfg.istrain) + elif self.cfg.BN_type == "IN": + x = self.instance_norm(x, name="%s_IN" % name) + else: + raise NotImplementedError('[ERROR] BN type [%s] is not implemented!' % self.cfg.BN_type) + x = self.activation(x, act_type=act_type) + x = tf.nn.conv2d(x, filter=weight, strides=[1, stride, stride, 1], padding=padding) + x = tf.nn.bias_add(x, bias) + return x + + def DeConvBlock(self, x, in_channels, out_channels, kernel_size, stride=1, name="DeConvBlock", + BN=True, use_bias=True, padding="VALID", act_type="relu", mode="CNA"): + assert (mode in ['CNA', 'NAC']), '[ERROR] Wrong mode in [%s]!' % sys.modules[__name__] + b, h, w, c = x.get_shape().as_list() + out_shape = [b, h * self.cfg.scale, w * self.cfg.scale, out_channels] + weight, bias = self.init_params(kernel_size=kernel_size, in_channels=out_channels, + out_channels=in_channels, use_bias=use_bias, name=name) + if mode == "CNA": + x = tf.nn.conv2d_transpose(x, filter=weight, output_shape=out_shape, + strides=[1, stride, stride, 1], padding=padding) + x = tf.nn.bias_add(x, bias) + if BN: + if self.cfg.BN_type == "BN": + x = self.batch_norm(x, training=True) + elif self.cfg.BN_type == "IN": + x = self.instance_norm(x, name="%s_IN" % name) + else: + raise NotImplementedError('[ERROR] BN type [%s] is not implemented!' % self.cfg.BN_type) + x = self.activation(x, act_type=act_type) + return x + elif mode == "NAC": + if BN: + if self.cfg.BN_type == "BN": + x = self.batch_norm(x, training=True) + elif self.cfg.BN_type == "IN": + x = self.instance_norm(x, name="%s_IN" % name) + else: + raise NotImplementedError('[ERROR] BN type [%s] is not implemented!' % self.cfg.BN_type) + x = self.activation(x, act_type=act_type) + x = tf.nn.conv2d_transpose(x, filter=weight, output_shape=out_shape, + strides=[1, stride, stride, 1], padding=padding) + x = tf.nn.bias_add(x, bias) + return x diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/LICENSE b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/LICENSE new file mode 100644 index 0000000000000000000000000000000000000000..5ea8a5f7b6ae91ebb12b7f2fa71a5432bb89de63 --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/LICENSE @@ -0,0 +1,284 @@ + Apache License + Version 2.0, January 2004 + http://www.apache.org/licenses/ + + TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION + + 1. Definitions. + + "License" shall mean the terms and conditions for use, reproduction, + and distribution as defined by Sections 1 through 9 of this document. + + "Licensor" shall mean the copyright owner or entity authorized by + the copyright owner that is granting the License. + + "Legal Entity" shall mean the union of the acting entity and all + other entities that control, are controlled by, or are under common + control with that entity. For the purposes of this definition, + "control" means (i) the power, direct or indirect, to cause the + direction or management of such entity, whether by contract or + otherwise, or (ii) ownership of fifty percent (50%) or more of the + outstanding shares, or (iii) beneficial ownership of such entity. + + "You" (or "Your") shall mean an individual or Legal Entity + exercising permissions granted by this License. + + "Source" form shall mean the preferred form for making modifications, + including but not limited to software source code, documentation + source, and configuration files. + + "Object" form shall mean any form resulting from mechanical + transformation or translation of a Source form, including but + not limited to compiled object code, generated documentation, + and conversions to other media types. + + "Work" shall mean the work of authorship, whether in Source or + Object form, made available under the License, as indicated by a + copyright notice that is included in or attached to the work + (an example is provided in the Appendix below). + + "Derivative Works" shall mean any work, whether in Source or Object + form, that is based on (or derived from) the Work and for which the + editorial revisions, annotations, elaborations, or other modifications + represent, as a whole, an original work of authorship. For the purposes + of this License, Derivative Works shall not include works that remain + separable from, or merely link (or bind by name) to the interfaces of, + the Work and Derivative Works thereof. + + "Contribution" shall mean any work of authorship, including + the original version of the Work and any modifications or additions + to that Work or Derivative Works thereof, that is intentionally + submitted to Licensor for inclusion in the Work by the copyright owner + or by an individual or Legal Entity authorized to submit on behalf of + the copyright owner. For the purposes of this definition, "submitted" + means any form of electronic, verbal, or written communication sent + to the Licensor or its representatives, including but not limited to + communication on electronic mailing lists, source code control systems, + and issue tracking systems that are managed by, or on behalf of, the + Licensor for the purpose of discussing and improving the Work, but + excluding communication that is conspicuously marked or otherwise + designated in writing by the copyright owner as "Not a Contribution." + + "Contributor" shall mean Licensor and any individual or Legal Entity + on behalf of whom a Contribution has been received by Licensor and + subsequently incorporated within the Work. + + 2. Grant of Copyright License. Subject to the terms and conditions of + this License, each Contributor hereby grants to You a perpetual, + worldwide, non-exclusive, no-charge, royalty-free, irrevocable + copyright license to reproduce, prepare Derivative Works of, + publicly display, publicly perform, sublicense, and distribute the + Work and such Derivative Works in Source or Object form. + + 3. Grant of Patent License. Subject to the terms and conditions of + this License, each Contributor hereby grants to You a perpetual, + worldwide, non-exclusive, no-charge, royalty-free, irrevocable + (except as stated in this section) patent license to make, have made, + use, offer to sell, sell, import, and otherwise transfer the Work, + where such license applies only to those patent claims licensable + by such Contributor that are necessarily infringed by their + Contribution(s) alone or by combination of their Contribution(s) + with the Work to which such Contribution(s) was submitted. If You + institute patent litigation against any entity (including a + cross-claim or counterclaim in a lawsuit) alleging that the Work + or a Contribution incorporated within the Work constitutes direct + or contributory patent infringement, then any patent licenses + granted to You under this License for that Work shall terminate + as of the date such litigation is filed. + + 4. Redistribution. You may reproduce and distribute copies of the + Work or Derivative Works thereof in any medium, with or without + modifications, and in Source or Object form, provided that You + meet the following conditions: + + (a) You must give any other recipients of the Work or + Derivative Works a copy of this License; and + + (b) You must cause any modified files to carry prominent notices + stating that You changed the files; and + + (c) You must retain, in the Source form of any Derivative Works + that You distribute, all copyright, patent, trademark, and + attribution notices from the Source form of the Work, + excluding those notices that do not pertain to any part of + the Derivative Works; and + + (d) If the Work includes a "NOTICE" text file as part of its + distribution, then any Derivative Works that You distribute must + include a readable copy of the attribution notices contained + within such NOTICE file, excluding those notices that do not + pertain to any part of the Derivative Works, in at least one + of the following places: within a NOTICE text file distributed + as part of the Derivative Works; within the Source form or + documentation, if provided along with the Derivative Works; or, + within a display generated by the Derivative Works, if and + wherever such third-party notices normally appear. The contents + of the NOTICE file are for informational purposes only and + do not modify the License. You may add Your own attribution + notices within Derivative Works that You distribute, alongside + or as an addendum to the NOTICE text from the Work, provided + that such additional attribution notices cannot be construed + as modifying the License. + + You may add Your own copyright statement to Your modifications and + may provide additional or different license terms and conditions + for use, reproduction, or distribution of Your modifications, or + for any such Derivative Works as a whole, provided Your use, + reproduction, and distribution of the Work otherwise complies with + the conditions stated in this License. + + 5. Submission of Contributions. Unless You explicitly state otherwise, + any Contribution intentionally submitted for inclusion in the Work + by You to the Licensor shall be under the terms and conditions of + this License, without any additional terms or conditions. + Notwithstanding the above, nothing herein shall supersede or modify + the terms of any separate license agreement you may have executed + with Licensor regarding such Contributions. + + 6. Trademarks. This License does not grant permission to use the trade + names, trademarks, service marks, or product names of the Licensor, + except as required for reasonable and customary use in describing the + origin of the Work and reproducing the content of the NOTICE file. + + 7. Disclaimer of Warranty. Unless required by applicable law or + agreed to in writing, Licensor provides the Work (and each + Contributor provides its Contributions) on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or + implied, including, without limitation, any warranties or conditions + of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A + PARTICULAR PURPOSE. You are solely responsible for determining the + appropriateness of using or redistributing the Work and assume any + risks associated with Your exercise of permissions under this License. + + 8. Limitation of Liability. In no event and under no legal theory, + whether in tort (including negligence), contract, or otherwise, + unless required by applicable law (such as deliberate and grossly + negligent acts) or agreed to in writing, shall any Contributor be + liable to You for damages, including any direct, indirect, special, + incidental, or consequential damages of any character arising as a + result of this License or out of the use or inability to use the + Work (including but not limited to damages for loss of goodwill, + work stoppage, computer failure or malfunction, or any and all + other commercial damages or losses), even if such Contributor + has been advised of the possibility of such damages. + + 9. Accepting Warranty or Additional Liability. While redistributing + the Work or Derivative Works thereof, You may choose to offer, + and charge a fee for, acceptance of support, warranty, indemnity, + or other liability obligations and/or rights consistent with this + License. However, in accepting such obligations, You may act only + on Your own behalf and on Your sole responsibility, not on behalf + of any other Contributor, and only if You agree to indemnify, + defend, and hold each Contributor harmless for any liability + incurred by, or claims asserted against, such Contributor by reason + of your accepting any such warranty or additional liability. + + END OF TERMS AND CONDITIONS + + APPENDIX: How to apply the Apache License to your work. + + To apply the Apache License to your work, attach the following + boilerplate notice, with the fields enclosed by brackets "[]" + replaced with your own identifying information. (Don't include + the brackets!) The text should be enclosed in the appropriate + comment syntax for the file format. We also recommend that a + file or class name and description of purpose be included on the + same "printed page" as the copyright notice for easier + identification within third-party archives. + + Copyright [yyyy] [name of copyright owner] + + 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. + +------------------ +Files: third_party/compute_library/... + +MIT License + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. + +------------------ +Files: ACKNOWLEDGEMENTS +LICENSE + +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + +1. Redistributions of source code must retain the above copyright notice, this + list of conditions and the following disclaimer. + +2. Redistributions in binary form must reproduce the above copyright notice, + this list of conditions and the following disclaimer in the documentation + and/or other materials provided with the distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND + ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED + WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE + DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR + ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES + (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; + LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND + ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS + SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +------------------ +Files: third_party/hexagon + +Copyright (c) 2016-2019, The Linux Foundation. All rights reserved. + +Redistribution and use in source and binary forms, with or without +modification, are permitted (subject to the limitations in the +disclaimer below) provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above + copyright notice, this list of conditions and the following + disclaimer in the documentation and/or other materials provided + with the distribution. + + * Neither the name of The Linux Foundation nor the names of its + contributors may be used to endorse or promote products derived + from this software without specific prior written permission. + +NO EXPRESS OR IMPLIED LICENSES TO ANY PARTY'S PATENT RIGHTS ARE +GRANTED BY THIS LICENSE. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT +HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED +WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF +MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. +IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR +ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE +GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER +IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR +OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN +IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. \ No newline at end of file diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/PreProcess.py b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/PreProcess.py new file mode 100644 index 0000000000000000000000000000000000000000..5ed011cdbf6684e72110a05a5a221ee64ca3f528 --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/PreProcess.py @@ -0,0 +1,98 @@ +# Copyright 2017 The TensorFlow 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. +# ============================================================================ +# Copyright 2021 Huawei Technologies Co., Ltd +# +# 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 npu_bridge.npu_init import * +import cv2 +import numpy as np +import random +from skimage import util + + +def add_noise(img): + mode_types = ['gaussian', 'localvar', 'poisson', 'speckle'] # 'salt', 'pepper', 's&p' is too fake + inx = int(np.random.choice(np.arange(len(mode_types)), 1)) + inx = 0 + img = cv2.cvtColor(img.astype(np.uint8), cv2.COLOR_BGR2RGB) + mean = random.random() * 0.001 # + 0.001#random.random() generates number between 0 and 1 + var = random.random() * 0.002 # + 0.01 + noise_img = util.random_noise(img.copy(), mode=mode_types[inx], + mean=mean, + var=var) + return noise_img + + +def augment_data(img_patch, flip, rot): # img_patchs : n,h,w,c + if flip==1: + img_patch = img_patch[:, ::-1, :] # hflip + elif flip==2: + img_patch = img_patch[::-1, :, :] # vflip + if rot==1: + img_patch = cv2.rotate(img_patch, cv2.ROTATE_90_CLOCKWISE) + elif rot==2: + img_patch = cv2.rotate(img_patch, cv2.ROTATE_90_COUNTERCLOCKWISE) + return img_patch + +def preprocess(imgs, cfg): + LR_patchs, HR_patchs = [], [] + for img in imgs: + + HR = cv2.imread(img.strip(), cv2.IMREAD_COLOR) + HR = (HR - 127.5) / 128 + h, w, c = HR.shape + + x_stride = w // (cfg.imagesize * cfg.scale) + y_stride = h // (cfg.imagesize * cfg.scale) + + for x in range(x_stride): + for y in range(y_stride): + HR_patch = HR[y * cfg.imagesize * cfg.scale:(y + 1) * cfg.imagesize * cfg.scale, + x * cfg.imagesize * cfg.scale:(x + 1) * cfg.imagesize * cfg.scale, :] + # add noise && add blur + t = np.random.randint(0, 2, 1) + if t == 0: + LR_patch = cv2.resize(HR_patch, dsize=None, fx=1 / cfg.scale, fy=1 / cfg.scale, + interpolation=cv2.INTER_CUBIC) + LR_patch = np.clip(LR_patch, -1.0, 1.0) + #LR_patch = add_noise(LR_patch) + else: + #LR_patch = add_noise(HR_patch) # [-1, 1] + LR_patch = cv2.resize(HR_patch, dsize=None, fx=1 / cfg.scale, + fy=1 / cfg.scale, interpolation=cv2.INTER_LINEAR) + # data augment + if cfg.istrain: + rot = np.random.randint(0, 3, 1) + flip = np.random.randint(0, 3, 1) + LR_patch = augment_data(LR_patch, flip, rot) + HR_patch = augment_data(HR_patch, flip, rot) + LR_patchs.append(LR_patch) + HR_patchs.append(HR_patch) + + return HR_patchs, LR_patchs + + + diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/README.md b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/README.md new file mode 100644 index 0000000000000000000000000000000000000000..edf8fc968564ffc42699a6382b0938b36b51c4fd --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/README.md @@ -0,0 +1,116 @@ +- [基本信息](#基本信息.md) +- [概述](#概述.md) +- [训练环境准备](#训练环境准备.md) +- [快速上手](#快速上手.md) +- [训练结果](#训练结果.md) +- [高级参考](#高级参考.md) +

基本信息

+ +**发布者（Publisher）：Huawei** + +**应用领域（Application Domain）：Computer Vision** + +**修改时间（Modified） ：2022.11.6** + +**框架（Framework）：TensorFlow 1.15.0** + +**描述（Description）：基于TensorFlow框架对高清图片重建相应的超分辨率图片的训练代码** + +

概述

+ +``` +SRFBN是采取反馈连接来提高重建超分辨率图片效果的网络模型 +``` +- 参考论文： + + https://arxiv.org/pdf/1903.09814.pdf + +- 参考实现： + + https://github.com/turboLIU/SRFBN-tensorflow/blob/master/train.py + +## 默认配置 + +- 训练数据集预处理： + + - 图像的输入尺寸为64*64 +- 测试数据集预处理： + + - 图像的输入尺寸为64*64 +- 训练超参 + + - Batch size： 1 + - Train epoch: 1000 + + +

快速上手

+ +- 数据集准备 +1. 模型训练使用DIV2K数据集。 + +## 模型训练 + +- 单卡训练 + + 1. 配置训练参数。 + + 首先在脚本test/train_performance_1p.sh中，配置batch_size、epochs、data_path等参数，请用户根据实际路径配置data_path，或者在启动训练的命令行中以参数形式下发。 + + ``` + batch_size=1 + epochs=1000 + data_path="../DIV2K/DIV2K_train_HR" + ``` + + 2. 启动训练。 + + 启动单卡训练 （脚本为SRFBN_for_TensorFlow/test/train_performance_1p.sh） + + ``` + bash train_performance_1p.sh --data_path=../DIV2K/DIV2K_train_HR + ``` + +

训练结果

+ +- 精度结果比对 + +| 精度指标项 | GPU实测 | NPU实测 | +| ---------- | ----------- | ----------- | +| PSNR | 6.706763287 | 5.831956861 | + +- 性能结果比对 + +| 性能指标项 | GPU实测 | NPU实测 | +| ---------- | -------------- | -------------- | +| FPS | 3.358950029841 | 4.976489075014 | + + +

高级参考

+ +## 脚本和示例代码 + +``` +├── Basic_Model.py //基本模型代码 +├── README.md //代码说明文档 +├── config.py //模型配置代码 +├── PreProcess.py //数据预处理代码 +├── requirements.txt //训练python依赖列表 +├── SRFBN_model.py //SRFBN网络模型代码 +├── test.py //测试代码 +├── train.py //训练代码 +├── test +│ ├──train_performance_1p.sh //单卡训练验证性能启动脚本 +│ ├──train_full_1p.sh //单卡全量训练启动脚本 +``` + +## 脚本参数 + +``` +--data_path 数据集路径，默认：path/data +--batch_size 每个NPU的batch size，默认：1 +--epochs 训练epcoh数量，默认：1000 +``` + +## 训练过程 + +1. 通过“模型训练”中的训练指令启动单卡训练。 \ No newline at end of file diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/SRFBN_model.py b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/SRFBN_model.py new file mode 100644 index 0000000000000000000000000000000000000000..4afaf9bb579f3c760394d28d114522832aa99c29 --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/SRFBN_model.py @@ -0,0 +1,209 @@ +# Copyright 2017 The TensorFlow 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. +# ============================================================================ +# Copyright 2021 Huawei Technologies Co., Ltd +# +# 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 npu_bridge.npu_init import * +import tensorflow as tf +import os +from Basic_Model import basic_network + + +class SRFBN(basic_network): + def __init__(self, sess, cfg): + super(SRFBN, self).__init__(cfg) + self.sess = sess + imageshape = [cfg.batchsize, cfg.imagesize, cfg.imagesize, cfg.c_dim] + labelshape = [cfg.batchsize, cfg.imagesize * cfg.scale, cfg.imagesize * cfg.scale, cfg.c_dim] + self.imageplaceholder = tf.placeholder(dtype=tf.float32, shape=imageshape, name="image") + self.labelplaceholder = tf.placeholder(dtype=tf.float32, shape=labelshape, name="label") + self.last_hidden = None + self.should_reset = True + self.outs = [] + #FB block that forwardpros feedback information + def FeedBackBlock(self, x, num_features, num_groups, act_type, name="FBB"): + if self.cfg.scale == 1: + stride = 1 + padding = "SAME" + kernel_size = 5 + if self.cfg.scale==2: + stride = 2 + padding = "SAME" + kernel_size = 6 + if self.cfg.scale == 3: + stride = 3 + padding = "SAME" + kernel_size = 7 + if self.cfg.scale == 4: + stride = 4 + padding = "SAME" + kernel_size = 8 + if self.should_reset: + self.last_hidden = x + self.should_reset = False + x = tf.concat([x, self.last_hidden], 3) + x = self.ConvBlock(x, 2*num_features, num_features, kernel_size=1, name="FeedBack_compress_in", + act_type=act_type) + + lr_features = [] + hr_features = [] + lr_features.append(x) + for i in range(num_groups): + x = tf.concat(lr_features, 3) + if i > 0: + x = self.ConvBlock(x, num_features*(i+1), num_features, kernel_size=1,stride=1, + padding=padding, act_type=act_type, name="%s_%d"%(name, i)) + x = self.DeConvBlock(x, num_features, num_features, kernel_size=kernel_size, stride=stride, + padding=padding, act_type=act_type, name="%s_%d"%(name, i)) + hr_features.append(x) + x = tf.concat(hr_features, 3) + if i > 0: + x = self.ConvBlock(x, num_features*(i+1), num_features, kernel_size=1, stride=1, + padding=padding, act_type=act_type, name="%s_%d"%(name, i)) + x = self.ConvBlock(x, num_features, num_features, kernel_size=kernel_size, stride=stride, + padding=padding, act_type=act_type, name="%s_%d"%(name, i)) + lr_features.append(x) + del hr_features + + x = tf.concat(lr_features[1:], 3) + + x = self.ConvBlock(x, num_features*num_groups, num_features, kernel_size=1, + act_type=act_type, name="FeedBack_compress_out") + + self.last_hidden = x + + return x + + def build(self): + if self.cfg.scale == 2: + stride = 2 + padding = "SAME" + kernel_size = 6 + if self.cfg.scale == 3: + stride = 3 + padding = "SAME" + kernel_size = 7 + if self.cfg.scale == 4: + stride = 4 + padding = "SAME" + kernel_size = 8 + # x = self.sub_mean(self.imageplaceholder) # normalize + + _, height, width, _ = self.imageplaceholder.get_shape().as_list() + + inter_size = tf.constant([height*self.cfg.scale, width*self.cfg.scale]) + inter_res = tf.image.resize_images(self.imageplaceholder, inter_size) + # inter_res = self.imageplaceholder + + x = self.ConvBlock(self.imageplaceholder, self.cfg.in_channels, 4 * self.cfg.num_features, kernel_size=3, + act_type=self.cfg.act_type, padding="SAME", name="conv_in") + x = self.ConvBlock(x, 4*self.cfg.num_features, self.cfg.num_features, kernel_size=1, + act_type=self.cfg.act_type, padding="SAME", name="feat_in") + # outs = [] + for i in range(self.cfg.num_steps): + if i == 0: + self.should_reset=True + t = self.FeedBackBlock(x, self.cfg.num_features, self.cfg.num_groups, self.cfg.act_type, name="FBB_%d"%i) + t = self.DeConvBlock(t, self.cfg.num_features, self.cfg.num_features, kernel_size=kernel_size, + stride=stride, padding=padding, act_type="relu", name="out_%d"%i) + t = self.ConvBlock(t, self.cfg.num_features, self.cfg.out_channels, kernel_size=3, stride=1, + act_type="tanh", padding="SAME", name="conv_out") + t = inter_res + t + t = tf.clip_by_value(t, -1.0, 1.0) + # t = t + inter_res + # t = self.add_mean(t) + self.outs.append(t) + + def train_step(self): + self.build() + print("This Net has Params num is %f MB" % (self.params_count * 4 / 1024 / 1024)) # float32 + tf.summary.image("image/HR", self.labelplaceholder, max_outputs=1) + out = tf.add_n(self.outs)/self.cfg.num_steps + + tf.summary.image("image/SR", out, max_outputs=1) + tf.summary.image("image/LR", self.imageplaceholder, max_outputs=1) + + self.l2_regularization_loss = tf.reduce_sum(tf.get_collection("weights_l2_loss")) + + self.losses = [self.calc_loss(x=x, y=self.labelplaceholder, loss_type=self.cfg.loss_type) for x in self.outs] + self.losses = tf.reduce_sum(self.losses)/len(self.losses)/self.cfg.batchsize + self.l2_regularization_loss + + tf.summary.scalar('loss/total', self.losses) + tf.summary.scalar('loss/l2_loss', self.l2_regularization_loss) + + self.merged_summary = tf.summary.merge_all() + self.saver = tf.train.Saver(max_to_keep=1) + #loading ckpt + def load(self): + model_name = "SRFBN.model" + model_dir = "%s_%s_%s_%s_c%d_x%s" % ( + "SRFBN", self.cfg.num_features, self.cfg.num_steps, self.cfg.num_groups, self.cfg.c_dim, self.cfg.scale) + checkpoint_dir = os.path.join(self.cfg.checkpoint_dir, model_dir) + ckpt = tf.train.get_checkpoint_state(checkpoint_dir) + if ckpt and ckpt.model_checkpoint_path: + ckpt_path = str(ckpt.model_checkpoint_path) + self.saver.restore(self.sess, os.path.join(os.getcwd(), ckpt_path)) + step = int(os.path.basename(ckpt_path).split('-')[1]) + print("\nCheckpoint Loading Success! %s\n" % ckpt_path) + else: + step = 0 + print("\nCheckpoint Loading Failed! \n") + + return step + #save model + def save(self, step): + model_name = "SRFBN.model" + model_dir = "%s_%s_%s_%s_c%d_x%s" % \ + ("SRFBN", self.cfg.num_features, self.cfg.num_steps, + self.cfg.num_groups, self.cfg.c_dim, self.cfg.scale) + checkpoint_dir = os.path.join(self.cfg.checkpoint_dir, model_dir) + + if not os.path.exists(checkpoint_dir): + os.makedirs(checkpoint_dir) + + self.saver.save(self.sess, + os.path.join(checkpoint_dir, model_name), + global_step=step) + #test + def test(self, width, height): + self.cfg.batchsize = 1 + testshape = [self.cfg.batchsize, height, width, self.cfg.c_dim] + labelshape = [self.cfg.batchsize, height*self.cfg.scale, width*self.cfg.scale, self.cfg.c_dim] + self.imageplaceholder = tf.placeholder(dtype=tf.float32, shape=testshape) + self.labelplaceholder = tf.placeholder(dtype=tf.float32, shape=labelshape) + self.build() + # self.outs = [self.add_mean(x) for x in self.outs] + out = tf.add_n(self.outs)/self.cfg.num_steps + # out = tf.concat(self.outs, -1) + return out + + +if __name__ == '__main__': + from config import SRFBN_config as config + cfg = config() + sess = tf.Session(config=npu_config_proto()) + net = SRFBN(sess, cfg) + train_step = net.train_step() diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/config.py b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/config.py new file mode 100644 index 0000000000000000000000000000000000000000..391bbadccb78021c08332117159329c5551b6325 --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/config.py @@ -0,0 +1,81 @@ +# Copyright 2017 The TensorFlow 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. +# ============================================================================ +# Copyright 2021 Huawei Technologies Co., Ltd +# +# 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 npu_bridge.npu_init import * +import os + +class config: + def __init__(self): + self.batchsize = 1 + self.Process_num = 3 + self.maxsize = 200 + self.ngpu = 1 + self.imagesize = 64 + self.scale = 3 + self.epoch = 1000 + #create ckpt,log,result dir + self.checkpoint_dir = "./model" + if not os.path.exists(self.checkpoint_dir): + os.mkdir(self.checkpoint_dir) + self.log_dir = "./log" + if not os.path.exists(self.log_dir): + os.mkdir(self.log_dir) + self.result = "./result" + if not os.path.exists(self.result): + os.mkdir(self.result) + + + +class SRFBN_config(config): + def __init__(self): + super(SRFBN_config, self).__init__() + self.istrain = True#is train or is test + self.istest = not self.istrain + self.c_dim = 3 #color channel can train one-channel pic or RGB pic + self.in_channels = 3 + self.out_channels = 3 + self.num_features = 32#base number of filter + self.num_steps = 4# timestep + self.num_groups = 6#the number of projection group of FBB feedbackblock + self.BN = True + if self.BN: + self.BN_type = "BN" # "BN" # or "IN" + self.act_type = "prelu" #activation function + self.loss_type = "L2" + self.lr_steps = [150, 300, 550, 750]#iteration + self.lr_gama = 1 + self.learning_rate = 2e-7#learning rate + self.load_premodel = True + #create dir + self.srfbn_logdir = "%s/srfbn" % self.log_dir + if not os.path.exists(self.srfbn_logdir): + os.mkdir(self.srfbn_logdir) + self.srfbn_result = "%s/srfbn" % self.result + if not os.path.exists(self.srfbn_result): + os.mkdir(self.srfbn_result) + diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/modelzoo_level.txt b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/modelzoo_level.txt new file mode 100644 index 0000000000000000000000000000000000000000..9981888d4330a88ecfc05366e577609f83017194 --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/modelzoo_level.txt @@ -0,0 +1,6 @@ +GPUStatus:OK +NPUMigrationStatus:OK +FuncStatus:OK +PrecisionStatus:POK +AutoTune:OK +PerfStatus:PERFECT \ No newline at end of file diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/requirements.txt b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/requirements.txt new file mode 100644 index 0000000000000000000000000000000000000000..e04023578103b7c569e1fd39a65bc81d44632a6b --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/requirements.txt @@ -0,0 +1,4 @@ +tensorflow==1.15.0 +cv2 +numpy +scikit-image \ No newline at end of file diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test.py b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test.py new file mode 100644 index 0000000000000000000000000000000000000000..be8f5570f924ef2ef612b529b78e32cfafeb0977 --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test.py @@ -0,0 +1,87 @@ +# Copyright 2017 The TensorFlow 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. +# ============================================================================ +# Copyright 2021 Huawei Technologies Co., Ltd +# +# 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 npu_bridge.npu_init import * +import tensorflow as tf + +from SRFBN_model import SRFBN +from PreProcess import * +from skimage.metrics import peak_signal_noise_ratio as compare_psnr +from skimage.metrics import _structural_similarity + + + + +def test_SRFBN(image_lr,image_hr): + + #image + height, width, _ = image_lr.shape + global load_flag + global srfbn + global out + if load_flag == 0: + srfbn = SRFBN(sess, cfg) + out = srfbn.test(width, height) + tf.global_variables_initializer().run(session=sess) + srfbn.saver = tf.train.Saver(max_to_keep=1) + srfbn.load() + srfbn.l2_regularization_loss = tf.reduce_sum(tf.get_collection("weights_l2_loss")) + srfbn.losses = [srfbn.calc_loss(x=x, y=srfbn.labelplaceholder, loss_type=srfbn.cfg.loss_type) for x in + srfbn.outs] + srfbn.losses = tf.reduce_sum(srfbn.losses) / len(srfbn.losses) / srfbn.cfg.batchsize + srfbn.l2_regularization_loss + load_flag += 1 + #cv2.namedWindow("result", 0) + + img_hr = image_hr.reshape([1,height*srfbn.cfg.scale,width*srfbn.cfg.scale,3]) + img_lr = image_lr.reshape([1, height, width, 3]) + output,err,l2_loss = sess.run([out,srfbn.losses,srfbn.l2_regularization_loss], feed_dict={srfbn.imageplaceholder: img_lr,srfbn.labelplaceholder:img_hr}) + output = output[0] * 128 + 127.5 + img_hr = img_hr.reshape([height*srfbn.cfg.scale,width*srfbn.cfg.scale,3]) + PSNR = compare_psnr(output, img_hr, data_range=255) + ssim = _structural_similarity.structural_similarity(output, img_hr,win_size=11, data_range=255, multichannel=True) + print("loss:[%.8f], l2_loss:[%.8f], PSNR:[%.8f], SSIM:[%.8f]"%(err,l2_loss,PSNR,ssim)) + #cv2.imshow("result", np.uint8(output)) + #cv2.waitKey(0) + + + +if __name__ == '__main__': + sess = tf.Session(config=npu_config_proto()) + from config import SRFBN_config + cfg = SRFBN_config() + cfg.istest = True + cfg.istrain = False + image = "/home/TestUser08/BUAA/Resolution_2K/DIV2K/DIV2K_valid_HR/0801.png" + batch_label,batch_lrimage = preprocess([image,],cfg) + batch_lrimage = np.array(batch_lrimage) + batch_label = np.array(batch_label) + load_flag = 0 + for i in range(batch_label.shape[0]): + test_SRFBN(batch_lrimage[i],batch_label[i]) + srfbn.sess.close() + diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test/.keep b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test/.keep new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test/train_full_1p.sh b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test/train_full_1p.sh new file mode 100644 index 0000000000000000000000000000000000000000..1242fb044895a27e42122ecc43b0f3950dd10619 --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test/train_full_1p.sh @@ -0,0 +1,184 @@ +#!/bin/bash + +########################################################## +#########第3行 至 100行，请一定不要、不要、不要修改########## +#########第3行 至 100行，请一定不要、不要、不要修改########## +#########第3行 至 100行，请一定不要、不要、不要修改########## +########################################################## +# shell脚本所在路径 +cur_path=`echo $(cd $(dirname $0);pwd)` + +# 判断当前shell是否是performance +perf_flag=`echo $0 | grep performance | wc -l` + +# 当前执行网络的名称 +Network=`echo $(cd $(dirname $0);pwd) | awk -F"/" '{print $(NF-1)}'` + +export RANK_SIZE=1 +export RANK_ID=0 +export JOB_ID=10087 + +# 路径参数初始化 +data_path="" +output_path="" + +# 帮助信息，不需要修改 +if [[ $1 == --help || $1 == -h ]];then + echo "usage:./train_performance_1P.sh " + echo " " + echo "parameter explain: + --data_path # dataset of training + --output_path # output of training + --train_steps # max_step for training + --train_epochs # max_epoch for training + --batch_size # batch size + -h/--help show help message + " + exit 1 +fi + +# 参数校验，不需要修改 +for para in $* +do + if [[ $para == --data_path* ]];then + data_path=`echo ${para#*=}` + elif [[ $para == --output_path* ]];then + output_path=`echo ${para#*=}` + elif [[ $para == --train_steps* ]];then + train_steps=`echo ${para#*=}` + elif [[ $para == --train_epochs* ]];then + train_epochs=`echo ${para#*=}` + elif [[ $para == --batch_size* ]];then + batch_size=`echo ${para#*=}` + fi +done + +# 校验是否传入data_path,不需要修改 +if [[ $data_path == "" ]];then + echo "[Error] para \"data_path\" must be config" + exit 1 +fi + +# 校验是否传入output_path,不需要修改 +if [[ $output_path == "" ]];then + output_path="./output/${ASCEND_DEVICE_ID}" +fi + +# 设置打屏日志文件名，请保留，文件名为${print_log} +print_log="./test/output/${ASCEND_DEVICE_ID}/train_${ASCEND_DEVICE_ID}.log" +modelarts_flag=`cat /etc/passwd |grep ma-user` +if [ x"${modelarts_flag}" != x ]; +then + echo "running with modelarts_flag..." + print_log_name=`ls /home/ma-user/modelarts/log/ | grep proc-rank` + print_log="/home/ma-user/modelarts/log/${print_log_name}" +fi +echo "### get your log here : ${print_log}" + +CaseName="" +function get_casename() +{ + if [ x"${perf_flag}" = x1 ]; + then + CaseName=${Network}_bs${batch_size}_${RANK_SIZE}'p'_'perf' + else + CaseName=${Network}_bs${batch_size}_${RANK_SIZE}'p'_'acc' + fi +} + +# 跳转到code目录 +cd ${cur_path}/../ +rm -rf ./test/output/${ASCEND_DEVICE_ID} +mkdir -p ./test/output/${ASCEND_DEVICE_ID} + +# 训练开始时间记录，不需要修改 +start_time=$(date +%s) +########################################################## +#########第3行 至 100行，请一定不要、不要、不要修改########## +#########第3行 至 100行，请一定不要、不要、不要修改########## +#########第3行 至 100行，请一定不要、不要、不要修改########## +########################################################## + +#========================================================= +#========================================================= +#========训练执行命令，需要根据您的网络进行修改============== +#========================================================= +#========================================================= +# 基础参数，需要模型审视修改 +# 您的训练数据集在${data_path}路径下，请直接使用这个变量获取 +# 您的训练输出目录在${output_path}路径下，请直接使用这个变量获取 +# 您的其他基础参数，可以自定义增加，但是batch_size请保留，并且设置正确的值 +batch_size=1 + +if [ x"${modelarts_flag}" != x ]; +then + python3 ./train.py +else + python3.7 ./train.py --data_path=${data_path} --output_path=${output_path} 1>${print_log} 2>&1 +fi + +# 性能相关数据计算 +StepTime=`((cat ${print_log} | grep "time" | head -n 1) && (cat ${print_log} | grep "time" | tail -n 1)) | awk -F ':' '{print $5 $6 }' | awk -F ',' '{print $1 $2}' | awk -F ' ' '{print $1;print $3}' | awk '{if (NR == 1){a=$1} else if (NR == 2){b=$1} else if (NR == 3){c=$1} else if (NR == 4){d=$1}} END {print (d-b)/(c-a)}'` +FPS=`awk 'BEGIN{printf "%.2f\n", '${batch_size}'/'${StepTime}'}'` +#PSNR值计算 +PSNR=`cat ${print_log} | grep "time" | tail -n 10 | awk -F ',' '{print $8}' | awk -F ':' '{sum+=$2} END {print sum/NR}'` +# 提取所有loss打印信息 +grep "loss:" ${print_log} | awk -F "," '{print $6}' > ./test/output/${ASCEND_DEVICE_ID}/my_output_loss.txt + + +########################################################### +#########后面的所有内容请不要修改########################### +#########后面的所有内容请不要修改########################### +#########后面的所有内容请不要修改########################### +########################################################### + +# 判断本次执行是否正确使用Ascend NPU +tf_flag=`echo ${Network} | grep TensorFlow | wc -l` +use_npu_flag=`grep "The model has been compiled on the Ascend AI processor" ${print_log} | wc -l` +if [ x"${use_npu_flag}" == x0 -a x"${tf_flag}" == x1 ]; +then + echo "------------------ ERROR NOTICE START ------------------" + echo "ERROR, your task haven't used Ascend NPU, please check your npu Migration." + echo "------------------ ERROR NOTICE END------------------" +else + echo "------------------ INFO NOTICE START------------------" + echo "INFO, your task have used Ascend NPU, please check your result." + echo "------------------ INFO NOTICE END------------------" +fi + +# 获取最终的casename，请保留，case文件名为${CaseName} +get_casename + +# 重命名loss文件 +if [ -f ./test/output/${ASCEND_DEVICE_ID}/my_output_loss.txt ]; +then + mv ./test/output/${ASCEND_DEVICE_ID}/my_output_loss.txt ./test/output/${ASCEND_DEVICE_ID}/${CaseName}_loss.txt +fi + +# 训练端到端耗时 +end_time=$(date +%s) +e2e_time=$(( $end_time - $start_time )) + +echo "------------------ Final result ------------------" +# 输出性能FPS/单step耗时/端到端耗时 +echo "Final Performance images/sec : $FPS" +echo "Final Performance sec/step : $StepTime" +echo "E2E Training Duration sec : $e2e_time" +echo "PSNR : $PSNR" +# 输出训练精度 +#echo "Final Train Accuracy : ${train_accuracy}" + +# 最后一个迭代loss值，不需要修改 +ActualLoss=(`awk 'END {print $NF}' ./test/output/${ASCEND_DEVICE_ID}/my_output_loss.txt`) + +#关键信息打印到${CaseName}.log中，不需要修改 +echo "Network = ${Network}" > $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "RankSize = ${RANK_SIZE}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "BatchSize = ${batch_size}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "DeviceType = `uname -m`" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "CaseName = ${CaseName}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "ActualFPS = ${FPS}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "TrainingTime = ${StepTime}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "ActualLoss = ${ActualLoss}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "E2ETrainingTime = ${e2e_time}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +#echo "TrainAccuracy = ${train_accuracy}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test/train_performance_1p.sh b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test/train_performance_1p.sh new file mode 100644 index 0000000000000000000000000000000000000000..708a9d41cfb630d36853a63f0dbaf4aabd602fa7 --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/test/train_performance_1p.sh @@ -0,0 +1,152 @@ +#!/bin/bash + +########################################################## +#########第3行 至 100行，请一定不要、不要、不要修改########## +#########第3行 至 100行，请一定不要、不要、不要修改########## +#########第3行 至 100行，请一定不要、不要、不要修改########## +########################################################## +# shell脚本所在路径 +cur_path=`echo $(cd $(dirname $0);pwd)` + +# 判断当前shell是否是performance +perf_flag=`echo $0 | grep performance | wc -l` + +# 当前执行网络的名称 +Network=`echo $(cd $(dirname $0);pwd) | awk -F"/" '{print $(NF-1)}'` + +export RANK_SIZE=1 +export RANK_ID=0 +export JOB_ID=10087 + +# 路径参数初始化 +data_path="" +output_path="" + +# 帮助信息，不需要修改 +if [[ $1 == --help || $1 == -h ]];then + echo "usage:./train_performance_1P.sh " + echo " " + echo "parameter explain: + --data_path # dataset of training + --output_path # output of training + --train_steps # max_step for training + --train_epochs # max_epoch for training + --batch_size # batch size + -h/--help show help message + " + exit 1 +fi + +# 参数校验，不需要修改 +for para in $* +do + if [[ $para == --data_path* ]];then + data_path=`echo ${para#*=}` + elif [[ $para == --output_path* ]];then + output_path=`echo ${para#*=}` + elif [[ $para == --train_steps* ]];then + train_steps=`echo ${para#*=}` + elif [[ $para == --train_epochs* ]];then + train_epochs=`echo ${para#*=}` + elif [[ $para == --batch_size* ]];then + batch_size=`echo ${para#*=}` + fi +done + +# 校验是否传入data_path,不需要修改 +if [[ $data_path == "" ]];then + echo "[Error] para \"data_path\" must be config" + exit 1 +fi + +# 校验是否传入output_path,不需要修改 +if [[ $output_path == "" ]];then + output_path="./output/${ASCEND_DEVICE_ID}" +fi + +CaseName="" +function get_casename() +{ + if [ x"${perf_flag}" = x1 ]; + then + CaseName=${Network}_bs${batch_size}_${RANK_SIZE}'p'_'perf' + else + CaseName=${Network}_bs${batch_size}_${RANK_SIZE}'p'_'acc' + fi +} + +# 跳转到code目录 +cd ${cur_path}/../ +rm -rf ./test/output/${ASCEND_DEVICE_ID} +mkdir -p ./test/output/${ASCEND_DEVICE_ID} + +# 训练开始时间记录，不需要修改 +start_time=$(date +%s) +########################################################## +#########第3行 至 90行，请一定不要、不要、不要修改########## +#########第3行 至 90行，请一定不要、不要、不要修改########## +#########第3行 至 90行，请一定不要、不要、不要修改########## +########################################################## + +#========================================================= +#========================================================= +#========训练执行命令，需要根据您的网络进行修改============== +#========================================================= +#========================================================= +# 基础参数，需要模型审视修改 +# 您的训练数据集在${data_path}路径下，请直接使用这个变量获取 +# 您的训练输出目录在${output_path}路径下，请直接使用这个变量获取 +# 您的其他基础参数，可以自定义增加，但是batch_size请保留，并且设置正确的值 +train_epochs=1 +batch_size=1 +print_log="./test/output/${ASCEND_DEVICE_ID}/train_${ASCEND_DEVICE_ID}.log" +python3.7 ./train.py --data_path=${data_path} --output_path=${output_path} 1>${print_log} 2>&1 +StepTime=`((cat ${print_log} | grep "time" | head -n 1) && (cat ${print_log} | grep "time" | tail -n 1)) | awk -F ':' '{print $5 $6 }' | awk -F ',' '{print $1 $2}' | awk -F ' ' '{print $1;print $3}' | awk '{if (NR == 1){a=$1} else if (NR == 2){b=$1} else if (NR == 3){c=$1} else if (NR == 4){d=$1}} END {print (d-b)/(c-a)}'` +FPS=`awk 'BEGIN{printf "%.2f\n", '${batch_size}'/'${StepTime}'}'` +#PSNR值计算 +PSNR=`cat ${print_log} | grep "time" | tail -n 10 | awk -F ',' '{print $8}' | awk -F ':' '{sum+=$2} END {print sum/NR}'` +# 提取所有loss打印信息 +grep "loss:" ${print_log} | awk -F "," '{print $6}' > ./test/output/${ASCEND_DEVICE_ID}/my_output_loss.txt + + +########################################################### +#########后面的所有内容请不要修改########################### +#########后面的所有内容请不要修改########################### +#########后面的所有内容请不要修改########################### +########################################################### + +# 获取最终的casename，请保留，case文件名为${CaseName} +get_casename + +# 重命名loss文件 +if [ -f ./test/output/${ASCEND_DEVICE_ID}/my_output_loss.txt ]; +then + mv ./test/output/${ASCEND_DEVICE_ID}/my_output_loss.txt ./test/output/${ASCEND_DEVICE_ID}/${CaseName}_loss.txt +fi + +# 训练端到端耗时 +end_time=$(date +%s) +e2e_time=$(( $end_time - $start_time )) + +echo "------------------ Final result ------------------" +# 输出性能FPS/单step耗时/端到端耗时 +echo "Final Performance images/sec : $FPS" +echo "Final Performance sec/step : $StepTime" +echo "E2E Training Duration sec : $e2e_time" +echo "PSNR : $PSNR" +# 输出训练精度 +#echo "Final Train Accuracy : ${train_accuracy}" + +# 最后一个迭代loss值，不需要修改 +ActualLoss=(`awk 'END {print $NF}' ./test/output/${ASCEND_DEVICE_ID}/${CaseName}_loss.txt`) + +#关键信息打印到${CaseName}.log中，不需要修改 +echo "Network = ${Network}" > $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "RankSize = ${RANK_SIZE}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "BatchSize = ${batch_size}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "DeviceType = `uname -m`" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "CaseName = ${CaseName}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "ActualFPS = ${FPS}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "TrainingTime = ${StepTime}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "ActualLoss = ${ActualLoss}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log +echo "E2ETrainingTime = ${e2e_time}" >> $cur_path/output/$ASCEND_DEVICE_ID/${CaseName}.log diff --git a/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/train.py b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/train.py new file mode 100644 index 0000000000000000000000000000000000000000..46f8808e3325498fc0cb737f8c36a8d63fd66595 --- /dev/null +++ b/TensorFlow/contrib/cv/SRFBN_for_TensorFlow/train.py @@ -0,0 +1,150 @@ +# Copyright 2017 The TensorFlow 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. +# ============================================================================ +# Copyright 2021 Huawei Technologies Co., Ltd +# +# 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 npu_bridge.npu_init import * +from SRFBN_model import SRFBN +import tensorflow as tf +import time +from PreProcess import * +from skimage.metrics import peak_signal_noise_ratio as comparepsnr +from skimage.metrics import _structural_similarity + +def train_SRFBN(dataset, sess, cfg): + # start put data in queue + with tf.device('/cpu:0'): + step = tf.Variable(0, trainable=False) + srfbn = SRFBN(sess=sess, cfg=cfg) + srfbn.train_step() + out = tf.add_n(srfbn.outs) / srfbn.cfg.num_steps + ## build Optimizer + #make lr_rate different in different stages + boundaries = [len(dataset)*epoch//cfg.batchsize for epoch in cfg.lr_steps] + values = [cfg.learning_rate*(cfg.lr_gama**i) for i in range(len(cfg.lr_steps)+1)] + lr = tf.train.piecewise_constant(step, boundaries, values) + update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS) + optimizer = tf.train.AdamOptimizer(learning_rate=lr) + with tf.control_dependencies(update_ops): + gs_vs = optimizer.compute_gradients(srfbn.losses) + with tf.device('/cpu:0'): + train_op = optimizer.apply_gradients(grads_and_vars=gs_vs, global_step=step) + + tf.global_variables_initializer().run(session=sess) + + summary_writer = tf.summary.FileWriter(cfg.srfbn_logdir, srfbn.sess.graph) + #load model + if srfbn.cfg.load_premodel: + counter = srfbn.load() + else: + counter = 0 + time_ = time.time() + print("\nNow Start Training...\n") + global_step = 0 + for ep in range(cfg.epoch): + + #pick pic randomly + pic_idx = np.random.permutation(len(dataset)) + picid = 0 + #load five pics one time + for i in range(0,len(dataset),5): + index = [] + for j in range(5): + index.append(pic_idx[i+j]) + imgnames = [] + for pic in index: + imgnames.append(dataset[pic]) + picid += 5 + print(imgnames) + batch_labels, batch_images = preprocess(imgnames, cfg) + patch_idx = list(range(len(batch_labels))) + #make the number of pic chunk divided by batchsize + if len(patch_idx) % cfg.batchsize != 0: + patch_idx.extend(list(np.random.choice(patch_idx, + cfg.batchsize * ((len(patch_idx) // cfg.batchsize)+1) - len(patch_idx)))) + + + patch_idx = np.random.permutation(patch_idx) + + + iterations = len(patch_idx) // cfg.batchsize + + + for it in range(iterations): + + idx = list(patch_idx[it * cfg.batchsize: (it+1)* cfg.batchsize]) + + + patch_labels = np.array(batch_labels)[idx] + + patch_images = np.array(batch_images)[idx] + + + output,_, loss,l2_loss,= srfbn.sess.run([out,train_op, srfbn.losses,srfbn.l2_regularization_loss], + feed_dict={srfbn.imageplaceholder: patch_images, + srfbn.labelplaceholder: patch_labels}) + output = output[0] * 128 + 127.5 + img_hr = patch_labels.reshape([srfbn.cfg.imagesize * srfbn.cfg.scale, srfbn.cfg.imagesize * srfbn.cfg.scale, 3]) + PSNR = comparepsnr(output, img_hr, data_range=255) + ssim = _structural_similarity.structural_similarity(output, img_hr, win_size=11, data_range=255, + multichannel=True) + + if it % 10 == 0: + print("Epoch:%2d, pic:%d, step:%2d, global_step:%d, time :%4.4f, loss:%.8f, l2_loss:%.8f, PSNR:%.8f, SSIM:%.8f" % ( + (ep + 1),picid, it,global_step,time.time() - time_, loss,l2_loss,PSNR,ssim)) + if it % 100 == 0: + srfbn.save(counter) + summary_str = srfbn.sess.run(srfbn.merged_summary, + feed_dict={srfbn.imageplaceholder: patch_images, + srfbn.labelplaceholder: patch_labels}) + summary_writer.add_summary(summary_str, counter) + + global_step += 1 + counter += 1 + +#train +def train(*args, **kwargs): + data_dir = kwargs["data_dir"] + imgs = [os.path.join(data_dir,data) for data in os.listdir(data_dir)] + + + + sess = tf.compat.v1.Session(config=npu_config_proto()) + + ## build NetWork + from config import SRFBN_config + cfg = SRFBN_config() + datasetet = imgs + train_SRFBN(datasetet, sess, cfg) + + + +if __name__ == '__main__': + import os + data_dir = "/home/TestUser08/BUAA/output_npu_20221021153629/SRFBN-tensorflow_npu_20221021153629/Resolution_2K/DIV2K/DIV2K_train_HR" + train(data_dir=data_dir) +