Initial commit

Dockerfile

+FROM image.sourcefind.cn:5000/dcu/admin/base/pytorch:1.10.0-centos7.6-dtk-22.10-py37-latest
+COPY requirements.txt requirements.txt
+RUN source /opt/dtk-22.10/env.sh
+RUN cp /usr/share/zoneinfo/Asia/Shanghai /etc/localtime && echo 'Asia/Shanghai' >/etc/timezone 
+ENV LANG C.UTF-8
+RUN pip install -r requirements.txt -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com

README.md

 # FLAVR
-## 模型介绍
-FLAVR是一种用于视频插值的深度学习模型，可以通过插值技术将低帧率视频转换为高帧率视频。
+
+## 论文
+
+` FLAVR: Flow-Agnostic Video Representations for Fast Frame Interpolation `
+
+- https://arxiv.org/pdf/2012.08512.pdf
+
 ## 模型结构
-3D U-Net结构、encoder部分采用ResNet-3D，decoder部分采用3D TransConv，以及Spatio-Temporal Feature Gating
+ FLAVR模型是一个3D U-Net模型，通过拓展2D U-Net，将2D卷积替换为3D卷积而得到的。该模型能够更准确地对输入帧之间的时间动态进行建模，从而获得更好的插值质量。 
+
+![](D:\git上传前\flavr_pytorch\doc\arch_dia.png)
+
+## 算法原理
+
+  3D U-Net结构、encoder部分采用ResNet-3D，decoder部分采用3D TransConv，以及Spatio-Temporal Feature Gating ![](D:\git上传前\flavr_pytorch\doc\原理.png)
+
+## 环境配置
+
+### Docker（方法一）
+
+此处提供[光源](https://www.sourcefind.cn/#/service-details)拉取docker镜像
+
+```
+docker pull image.sourcefind.cn:5000/dcu/admin/base/pytorch:1.10.0-centos7.6-dtk-22.10-py37-latest
+docker run -it --network=host --name=flavr --privileged --device=/dev/kfd --device=/dev/dri --ipc=host --shm-size=32G  --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -u root --ulimit stack=-1:-1 --ulimit memlock=-1:-1 image.sourcefind.cn:5000/dcu/admin/base/pytorch:1.10.0-centos7.6-dtk-22.10-py37-latest
+```
+
+### Dockerfile（方法二）
+
+dockerfile使用方法
+
+```
+docker build --no-cache -t flavr:latest .
+docker run -dit --network=host --name=flavr --privileged --device=/dev/kfd --device=/dev/dri --ipc=host --shm-size=16G  --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -u root --ulimit stack=-1:-1 --ulimit memlock=-1:-1 flavr:latest
+docker exec -it flavr /bin/bash
+pip install -r requirements.txt
+```
+
+### Anaconda（方法三）
+
+关于本项目DCU显卡所需的特殊深度学习库可从[光合](https://developer.hpccube.com/tool/)开发者社区下载安装。
+
+```
+DTK驱动：dtk22.10
+python：python3.7
+torch==1.10.0a0+git2040069.dtk2210
+torchvision==0.10.0a0+e04d001.dtk2210
+```
+
+`Tips：以上dtk驱动、python等DCU相关工具版本需要严格一一对应`
+
+其它非深度学习库参照requirements.txt安装：
+
+```
+Fpip install -r requirements.txt
+```
 
 ## 数据集
-目前模型在Vimeo-90K 数据集上训练
 
-可以通过[此链接](http://toflow.csail.mit.edu/)进行数据下载
+`模型使用数据为 Vimeo-90K `
+
+-  http://toflow.csail.mit.edu/
+
+项目中已提供用于试验训练的迷你数据集，训练数据目录结构如下，用于正常训练的完整数据集请按此目录结构进行制备：
+
+```
+ ── datasets
+   │   ├── readme.txt
+   │   ├── sep_testlist.txt 
+   │   ├── sep_trainlist.txt 
+   │   ├── sequences
+          │    ├── xx/xx/xxx.png
+          │    ├── xx/xx/xxx.png
+```
+
+
+
+
+## 训练
+### 单机单卡
+
+```
+python main.py --batch_size 32 --test_batch_size 32 --dataset vimeo90K_septuplet --loss 1*L1 --max_epoch 200 --lr 0.0002 --data_root datasets --n_outputs 1 --num_gpu 1
+```
+
+### 单机多卡
 
-## 训练及推理
-### 环境配置
-python依赖安装：
+```
+python main.py --batch_size 32 --test_batch_size 32 --dataset vimeo90K_septuplet --loss 1*L1 --max_epoch 200 --lr 0.0002 --data_root datasets --n_outputs 1 --num_gpu 4
+```
 
-    Python==3.7.11
-    numpy==1.19.2
-    PyTorch ==1.10.0a0+git2040069.dtk2210
-### 训练
-训练命令：
+## 推理
 
-    python main.py --batch_size 32 \
-                   --test_batch_size 32 \
-                   --dataset vimeo90K_septuplet \
-                   --loss 1*L1 \
-                   --max_epoch 200 \
-                   --lr 0.0002 \
-                   --data_root <dataset_path> \
-                   --n_outputs 1
+    python test.py --dataset vimeo90K_septuplet --data_root <data_path> --load_from <saved_model> --n_outputs 1
 
-GoPro 数据集上的训练类似，更改`n_outputs`为 7 以进行 8 倍插值。
+## result
 
+测试图
 
+![](D:\git上传前\flavr_pytorch\doc\sprite.gif)
 
-## 性能和准确率数据
-测试数据：[test data](http://toflow.csail.mit.edu/)，使用的加速卡:2张 DCU。
+### 精度
 
+测试数据：http://toflow.csail.mit.edu/，使用的加速卡:Z100L。
+
+根据测试结果情况填写表格：
+
+|  FLAVR   |   PSNR    | SSIM     |   速度   |
+| :------: | :-------: | -------- | :------: |
+| vimeo-90 | 18.511020 | 0.702564 | 3.46it/s |
+
+## 应用场景
+
+### 算法类别
+
+`图像超分`
+
+### 热点应用行业
+
+`设计`
 
 ## 源码仓库及问题反馈
-* [https://github.com/tarun005/FLAVR](https://github.com/tarun005/FLAVR)
-## 参考
-* [https://github.com/tarun005/FLAVR](https://github.com/tarun005/FLAVR)
+
+*   https://github.com/tarun005/FLAVR 
+## 参考资料
+*  https://github.com/tarun005/FLAVR 
+

datasets/readme.txt

+========================================================================================
+=== Vimeo-90K Septuplet Dataset for Video Denoising, Deblocking, and Super-resolution
+========================================================================================
+
+Vimeo-90K septuplets dataset contains 91701 septuplets (each septuplets is a short RGB video sequence that consists of 7 frames) from 39k video clips with fixed resolution 448x256. This dataset is designed to video denoising, deblocking, and super-resolution. All the videos are downloaded from vimeo.com.
+
+========================================================================================
+=== Folder structure
+========================================================================================
+
+- sequences: This folder stores all 91701 septuplets. It uses a two-level folder structure, where each folder "%05d/%04d" contains a short video sequence consists of 7 frames: im1.png, im2.png, ..., im7.png.
+
+- sep_trainlist.txt: contains the list of sequences for training.
+
+- sep_testlist.txt: contains the list of sequences for test.
+
+======================================================
+=== Citation
+======================================================
+
+If you use this dataset in your work, please cite the following work:
+
+@article{xue17toflow,
+  author = {Xue, Tianfan and Chen, Baian and Wu, Jiajun and Wei, Donglai and Freeman, William T},
+  title = {Video Enhancement with Task-Oriented Flow},
+  journal = {arXiv},
+  year = {2017}
+}
+
+For questions, please contact Tianfan Xue (tianfan.xue@gmail.com), Baian Chen(baian@mit.edu), Jiajun Wu (jiajunwu@mit.edu), or Donglai Wei(donglai@csail.mit.edu).
+
+For more information, please refers to our project website and github repo:
+
+Project website: http://toflow.csail.mit.edu/
+Github repo: https://github.com/anchen1011/toflow
+
+======================================================
+=== Disclaimer
+======================================================
+
+This dataset is for non-commercial usage only.

datasets/sep_testlist.txt

+00001/0266
\ No newline at end of file

datasets/sep_trainlist.txt

+00001/0001
\ No newline at end of file