位置：首页-资讯-后端开发

Python DPED机器学习之实现照片美化

2024-04-02 19:55

短信预约 -IT技能 免费直播动态提醒

前言

最近发现了一个可以把照片美化的项目，自己玩了玩，挺有意思的，分享一下。

Github地址：DPED项目地址

下面来看看项目怎么玩？先放一些项目给出的效果图。可以看出照片更明亮好看了。

环境部署

项目结构

下面是项目的原始结构：

tensorflow安装

按照项目的说明，我们需要安装tensorflow以及一些必要的库。

如果安装gpu版本的tensorflow需要对照一下

tensorflow官方对照地址：TensorFlow官方CUDA版本对照

我的cuda是11.1的版本，按照tensorflow后还是缺少部分dll，如果有相同问题的，可以用我提供的资源包提取码：TUAN。

缺少哪个dll，直接复制到你的NVIDIA GPU Computing Toolkit目录对应cuda的bin目录下。

按照自己的版本来，我的tensorflow命令如下：


pip install tensorflow-gpu==2.4.2 -i https://pypi.douban.com/simple
pip install tf-nightly -i https://pypi.douban.com/simple

其他依赖安装

Pillow, scipy, numpy, imageio安装


pip install Pillow -i https://pypi.douban.com/simple
pip install scipy -i https://pypi.douban.com/simple
pip install numpy -i https://pypi.douban.com/simple
pip install imageio -i https://pypi.douban.com/simple

VGG-19下载

因为模型文件太大，github的项目中无法上传这么大的文件，作者让我们自己下。

我把DPED的资源包统一打包了，也可以从我的云盘下载，放到项目的vgg_pretrained目录下。下图是资源包的目录

资源包地址提取码：TUAN。

项目运行

项目需要的环境我们都装好了，我们跳过训练的部分，测试model的方法官方给出了命令。

准备图片素材

我准备了几张图，就不全展示了，展示其中的一张。

按照项目的要求，需要放在对应的目录下。

测试效果

执行命令


python test_model.py model=iphone_orig test_subset=full resolution=orig use_gpu=true

执行过程


(tensorflow) C:\Users\yi\PycharmProjects\DPED>python test_model.py model=iphone_orig test_subset=full resolution=orig use_gpu=true
2021-11-27 23:42:57.922965: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll
2021-11-27 23:43:00.532645: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to
use the following CPU instructions in performance-critical operations:  AVX2
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
2021-11-27 23:43:00.535946: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library nvcuda.dll
2021-11-27 23:43:00.559967: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1720] Found device 0 with properties:
pciBusID: 0000:01:00.0 name: GeForce GTX 1070 computeCapability: 6.1
coreClock: 1.759GHz coreCount: 15 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 238.66GiB/s
2021-11-27 23:43:00.560121: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll
2021-11-27 23:43:00.577706: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll
2021-11-27 23:43:00.577812: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll
2021-11-27 23:43:00.588560: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cufft64_10.dll
2021-11-27 23:43:00.591950: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library curand64_10.dll
2021-11-27 23:43:00.614412: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusolver64_10.dll
2021-11-27 23:43:00.624267: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusparse64_11.dll
2021-11-27 23:43:00.626309: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudnn64_8.dll
2021-11-27 23:43:00.626481: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1862] Adding visible gpu devices: 0
2021-11-27 23:43:01.112598: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1261] Device interconnect StreamExecutor with strength 1 edge matrix:
2021-11-27 23:43:01.112756: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1267]      0
2021-11-27 23:43:01.113098: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1280] 0:   N
2021-11-27 23:43:01.113463: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1406] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 6720 MB
 memory) -> physical GPU (device: 0, name: GeForce GTX 1070, pci bus id: 0000:01:00.0, compute capability: 6.1)
2021-11-27 23:43:01.114296: I tensorflow/compiler/jit/xla_gpu_device.cc:99] Not creating XLA devices, tf_xla_enable_xla_devices not set
WARNING:tensorflow:From C:\Users\yi\AppData\Roaming\Python\Python36\site-packages\tensorflow\python\compat\v2_compat.py:96: disable_resource_variables (from tensorflow.p
ython.ops.variable_scope) is deprecated and will be removed in a future version.
Instructions for updating:
non-resource variables are not supported in the long term
2021-11-27 23:43:01.478512: I tensorflow/compiler/jit/xla_cpu_device.cc:41] Not creating XLA devices, tf_xla_enable_xla_devices not set
2021-11-27 23:43:01.479339: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1720] Found device 0 with properties:
pciBusID: 0000:01:00.0 name: GeForce GTX 1070 computeCapability: 6.1
coreClock: 1.759GHz coreCount: 15 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 238.66GiB/s
2021-11-27 23:43:01.479747: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll
2021-11-27 23:43:01.480519: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll
2021-11-27 23:43:01.480927: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll
2021-11-27 23:43:01.481155: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cufft64_10.dll
2021-11-27 23:43:01.481568: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library curand64_10.dll
2021-11-27 23:43:01.481823: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusolver64_10.dll
2021-11-27 23:43:01.482188: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusparse64_11.dll
2021-11-27 23:43:01.482416: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudnn64_8.dll
2021-11-27 23:43:01.482638: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1862] Adding visible gpu devices: 0
2021-11-27 23:43:01.482959: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1720] Found device 0 with properties:
pciBusID: 0000:01:00.0 name: GeForce GTX 1070 computeCapability: 6.1
coreClock: 1.759GHz coreCount: 15 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 238.66GiB/s
2021-11-27 23:43:01.483077: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll
2021-11-27 23:43:01.483254: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll
2021-11-27 23:43:01.483426: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll
2021-11-27 23:43:01.483638: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cufft64_10.dll
2021-11-27 23:43:01.483817: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library curand64_10.dll
2021-11-27 23:43:01.484052: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusolver64_10.dll
2021-11-27 23:43:01.484250: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusparse64_11.dll
2021-11-27 23:43:01.484433: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudnn64_8.dll
2021-11-27 23:43:01.484662: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1862] Adding visible gpu devices: 0
2021-11-27 23:43:01.484841: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1261] Device interconnect StreamExecutor with strength 1 edge matrix:
2021-11-27 23:43:01.484984: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1267]      0
2021-11-27 23:43:01.485152: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1280] 0:   N
2021-11-27 23:43:01.485395: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1406] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 6720 MB
 memory) -> physical GPU (device: 0, name: GeForce GTX 1070, pci bus id: 0000:01:00.0, compute capability: 6.1)
2021-11-27 23:43:01.485565: I tensorflow/compiler/jit/xla_gpu_device.cc:99] Not creating XLA devices, tf_xla_enable_xla_devices not set
2021-11-27 23:43:01.518135: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:196] None of the MLIR optimization passes are enabled (registered 0 passes)
Testing original iphone model, processing image 3.jpg
2021-11-27 23:43:01.863678: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudnn64_8.dll
2021-11-27 23:43:02.517063: I tensorflow/core/platform/windows/subprocess.cc:308] SubProcess ended with return code: 0
 
2021-11-27 23:43:02.632790: I tensorflow/core/platform/windows/subprocess.cc:308] SubProcess ended with return code: 0
 
2021-11-27 23:43:03.210892: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll
2021-11-27 23:43:03.509052: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll
Lossy conversion from float32 to uint8. Range [-0.06221151351928711, 1.0705437660217285]. Convert image to uint8 prior to saving to suppress this warning.
Lossy conversion from float32 to uint8. Range [-0.06221151351928711, 1.0705437660217285]. Convert image to uint8 prior to saving to suppress this warning.
Testing original iphone model, processing image 4.jpg
Lossy conversion from float32 to uint8. Range [-0.05176264047622681, 1.0500218868255615]. Convert image to uint8 prior to saving to suppress this warning.
Lossy conversion from float32 to uint8. Range [-0.05176264047622681, 1.0500218868255615]. Convert image to uint8 prior to saving to suppress this warning.
Testing original iphone model, processing image 5.jpg
Lossy conversion from float32 to uint8. Range [-0.03344374895095825, 1.0417983531951904]. Convert image to uint8 prior to saving to suppress this warning.
Lossy conversion from float32 to uint8. Range [-0.03344374895095825, 1.0417983531951904]. Convert image to uint8 prior to saving to suppress this warning.
Testing original iphone model, processing image 6.jpg
Lossy conversion from float32 to uint8. Range [-0.03614246845245361, 1.063475251197815]. Convert image to uint8 prior to saving to suppress this warning.
Lossy conversion from float32 to uint8. Range [-0.03614246845245361, 1.063475251197815]. Convert image to uint8 prior to saving to suppress this warning.

项目会生成前后对比图以及最终结果图。

前后效果图，左边为原始图，右边为对比图。