Category: provider
Model Studio AnimateAnyone
Validation
CODEBLOCK0
Pass criteria: command exits 0 and output/aliyun-animate-anyone/validate.txt is generated.
Output And Evidence
- - Save normalized request payloads, detection outputs, template IDs, and task polling snapshots under
output/aliyun-animate-anyone/. - Record whether the result should keep the reference image background or the source video background.
Use AnimateAnyone when the task needs motion transfer from a template video rather than plain talking-head animation.
Critical model names
Use these exact model strings:
- - INLINECODE2
- INLINECODE3
- INLINECODE4
Selection guidance:
- - Run image detection first.
- Run template generation on the source motion video.
- Use
animate-anyone-gen2 for the final video job.
Prerequisites
- - China mainland (Beijing) only.
- Set
DASHSCOPE_API_KEY in your environment, or add dashscope_api_key to ~/.alibabacloud/credentials. - Input files must be public HTTP/HTTPS URLs.
Normalized interface (video.animate_anyone)
Detect Request
- -
model (string, optional): default INLINECODE10 - INLINECODE11 (string, required)
Template Request
- -
model (string, optional): default INLINECODE13 - INLINECODE14 (string, required)
Generate Request
- -
model (string, optional): default INLINECODE16 - INLINECODE17 (string, required)
- INLINECODE18 (string, required)
- INLINECODE19 (bool, optional): whether to keep the input image background
Response
- -
task_id (string) - INLINECODE21 (string)
- INLINECODE22 (string, when finished)
Quick start
CODEBLOCK1
Operational guidance
- - The action template must come from the official template-generation API.
- Full-body images work best when
use_ref_img_bg=false; half-body images are not recommended in that mode. - This skill is best for dancing or large body motion transfer, not generic talking-head tasks.
Output location
- - Default output: INLINECODE24
- Override base dir with
OUTPUT_DIR.
References
技能名称: aliyun-animate-anyone
详细描述:
分类: provider
Model Studio AnimateAnyone
验证
bash
mkdir -p output/aliyun-animate-anyone
python -m pycompile skills/ai/video/aliyun-animate-anyone/scripts/prepareanimateanyonerequest.py && echo pycompileok > output/aliyun-animate-anyone/validate.txt
通过标准: 命令退出码为0且已生成 output/aliyun-animate-anyone/validate.txt。
输出与证据
- - 将标准化请求负载、检测输出、模板ID以及任务轮询快照保存在 output/aliyun-animate-anyone/ 目录下。
- 记录结果应保留参考图像背景还是源视频背景。
当任务需要从模板视频进行动作迁移，而非简单的说话头像动画时，使用AnimateAnyone。
关键模型名称
请使用以下精确的模型字符串：
- - animate-anyone-detect-gen2
- animate-anyone-template-gen2
- animate-anyone-gen2
选择指南：
- - 首先运行图像检测。
- 对源动作视频运行模板生成。
- 在最终视频任务中使用 animate-anyone-gen2。
前提条件
- - 仅限中国大陆（北京）区域。
- 在环境中设置 DASHSCOPEAPIKEY，或将 dashscopeapikey 添加到 ~/.alibabacloud/credentials 文件中。
- 输入文件必须是公开的HTTP/HTTPS URL。
标准化接口 (video.animate_anyone)
检测请求
- - model (字符串，可选): 默认为 animate-anyone-detect-gen2
- image_url (字符串，必填)
模板请求
- - model (字符串，可选): 默认为 animate-anyone-template-gen2
- video_url (字符串，必填)
生成请求
- - model (字符串，可选): 默认为 animate-anyone-gen2
- imageurl (字符串，必填)
- templateid (字符串，必填)
- userefimg_bg (布尔值，可选): 是否保留输入图像背景
响应
- - taskid (字符串)
- taskstatus (字符串)
- video_url (字符串，任务完成时返回)
快速开始
bash
python skills/ai/video/aliyun-animate-anyone/scripts/prepareanimateanyone_request.py \
--image-url https://example.com/dancer.png \
--template-id tmpl_xxx \
--use-ref-img-bg
操作指南
- - 动作模板必须来自官方模板生成API。
- 当 userefimg_bg=false 时，全身图像效果最佳；在该模式下不建议使用半身图像。
- 此技能最适合舞蹈或大幅身体动作迁移，不适用于通用说话头像任务。
输出位置
- - 默认输出: output/aliyun-animate-anyone/request.json
- 可通过 OUTPUT_DIR 覆盖基础目录。
参考资料
