From a1c23acc537e126bac9937bc01a66c440afa3425 Mon Sep 17 00:00:00 2001
From: FaceZhao <facezhao@163.com>
Date: Tue, 22 Aug 2023 03:36:34 +0000
Subject: [PATCH] add meta

---
 README.md | 43 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)

diff --git a/README.md b/README.md
index d8f40d7..c4cb020 100644
--- a/README.md
+++ b/README.md
@@ -1,3 +1,46 @@
+---
+backbone:
+- diffusion
+domain:
+- multi-modal
+frameworks:
+- pytorch
+license: CC-BY-NC-ND
+metrics:
+- realism
+- image-video similarity
+studios:
+- damo/Image-to-Video
+tags:
+- image2video generation
+- diffusion model
+- 图到视频
+- 图生视频
+- 图片生成视频
+- 生成
+tasks:
+- image-to-video
+widgets:
+  - examples:
+      - inputs:
+          - data: XXX/test.jpg.
+            name: image_path
+        name: 1
+        title: 示例1
+    inferencespec:
+      cpu: 4
+      gpu: 1
+      gpu_memory: 15G
+      memory: 32G
+    inputs:
+      - name: image_path
+        title: 图片的路径
+        type: str
+        validator:
+          max_words: /
+    task: image-to-video
+---
+
 # Image-to-Video
 
 本项目**MS-Image2Video**旨在解决根据输入图像生成高清视频任务。**MS-Image2Video**由达摩院研发的高清视频生成基础模型，其核心部分包含两个阶段，分别解决语义一致性和清晰度的问题，参数量共计约37亿，模型经过在大规模视频和图像数据混合预训练，并在少量精品数据上微调得到，该数据分布广泛、类别多样化，模型对不同的数据均有良好的泛化性。项目于现有的视频生成模型，**MS-Image2Video**在清晰度、质感、语义、时序连续性等方面均具有明显的优势。