Tell us what feature you want?

开发者定义openapi, 从openapi生成插件协议部分

What happened?

A bug happened!
17:47:12.411
terminate called after throwing an instance of 'spdlog::spdlog_ex'
17:47:12.411
what(): Failed opening file ./log/server/wrapper.log for writing: No such file or directory

Additional Information

No response

• AtheProxy
• AtheLoader
• AtheGate

Use protobuff to comunicate with the python inference

• 默认配置注入
• helm k8s 集成

add support for conda baseimage build.

Welcome update to OpenMMLab 2.0

I am Vansin, the technical operator of OpenMMLab. In September of last year, we announced the release of OpenMMLab 2.0 at the World Artificial Intelligence Conference in Shanghai. We invite you to upgrade your algorithm library to OpenMMLab 2.0 using MMEngine, which can be used for both research and commercial purposes. If you have any questions, please feel free to join us on the OpenMMLab Discord at https://discord.gg/amFNsyUBvm or add me on WeChat (van-sin) and I will invite you to the OpenMMLab WeChat group.

Here are the OpenMMLab 2.0 repos branches:

Attention: please create a new virtual environment for OpenMMLab 2.0.

What happened?

A bug happened!
Where there is no http_listen filed in aiges.toml, wrapper will not start.

invalied Configure in read s3fd0bd99=>http_listen
EngService exit

Additional Information

No response

Hi, wei.
This is an amazing and useful tool and I want to know can I use service mesh framework in this project?

Thx.

1. 升级至模型即服务是否考虑支持多模型编排
2. 支持python,cpp预后处理脚本仓库建设，支持onnx runtime推理框架 或自定义推理backend
3. python推理多进程支持 multiprocessing 执行exec

aiges's base image ,aiges_gpu image CI build and push to ecr registriy

• 部署文档优化，同步到开源门户
• webgate文档
• atoms文档
• loadbalance文档
• polaris文档
• asectl文档

What happened?

A bug happened!
if no such field in aiges.toml, service will not work normally.
[ws]
readTimeout = 15
sessionTimeout = 180

Additional Information

No response

need to trace

What happened?

最近使用——aiges- gpu:10.1-1.17-3.9.13-ubuntu1804-V3.0.5版本镜像，在验证py加载器时失败，同一个加载器在使用原来的老镜像aiges- gpu:10.1-1.17-3.9.13-ubuntu1804-V2.1.0时能够验证成功。
查看容器日志，可能是打开日志文件失败：Failed opening file ./log/server/wrapper.log for writing: No such file or directory，希望排查一下这种问题的原因

Additional Information

No response

as https://github.com/xfyun/aiges/blob/master/doc/design/wrapper_interface.md mentioned:

here is some sugguestions from the inner disscusions：

1. support multiple run tests with different test values.
2. support a template method to auto generate wrapper.py code project.
3. code structure need enhance.
4. mulit thread/process need be considered

1、用户通过在 wrapper.py中 编写推理逻辑，定义输入、输出数据、即可自动生成 schema文件，该文件可通过 ase_ctl 工具导入配置中心

goreportcard.com

need to port the xtest code

The aiges loader process cannot be closed with ctrl + c, and the kill command is also required. We have this problem inside the docker container.

Tell us what feature you want?

build xtest to docker

make the demo to submodule

Tell us what feature you want?

Define a yaml like this:

apiVersion: v1
kind: AIService
metadata:
  name: yolov5-demo-service
spec:
  containerSpec: ...
  modelSpec: ...

After defined that , we can apply this to create an API using kubectl command native.

Tell us what feature you want?

support stream processing

the dockerfile in mmocr is lost.

first..

can't pull release docker images. please check

• 在线文档系统CI CD
• AIGES 持续构建workflow
• AIGES 持续release workflow

What happened?

use command to init wrapper:
python3 -m aiges create -n mnist

wrapper.py
-->class UserRequest
---->input2 = StringBodyField(key="switch", value="ctrl")

ERROR: StringBodyField value must be bytes String...
input2 = StringBodyField(key="switch", value="ctrl")

Additional Information

No response

when user doesn't supply the function parameters, https://github.com/xfyun/aiges/blob/5364e44aaea2771c34991a2c97e4f5834f93fa90/ai_cpython_wrapper/wrapper.cpp#L151 will crashed

Tell us what feature you want?

Support Gradio , to generate gradio APP

Tell us what feature you want?

Although we already support websocket protocol, we need a stream implment like ws because of our longtime inference.

当研发一个新能力并对外发布一个新能力的时候，分别对应两阶段环境部署过程。
开发阶段: 往往需要setup一些框架以及框架周边工具，以及开发者自己定义的工具类, 等训练出模型；
推理阶段: 使用模型或者封装了模型的engine进行推理，模型根据不同的框架也存在不同的格式，但是业界也在逐步向onnx模型范式演进。这一阶段的环境依赖往往比较纯粹， 至少比开发阶段要少很多依赖。

问题以及痛点: 算法开发者当前同时也会推理服务开发，对大多数开发者，他们很难意识到2个环境的差异，因此往往在infer阶段共享使用了dev环境的依赖，造成infer阶段依赖特别冗余。再者由于python环境的不一致，pip freeze生成的一些依赖列表，往往无法自动化完成在新环境完成原环境的最小复现。

对此，我们再构建基础镜像时，也考虑了一些方案但是都并不完美，在发现 https://github.com/gaocegege 发起的 envd后，我们似乎发现有了同一目标项目，因此，我们计划后续在envd项目中来完善并解决这个问题

参考: https://github.com/xfyun/proposals/blob/main/athenaloader/asectl_design.md

开发 ase_ctl 工具

Tell us what feature you want?

we want to using openfaass framework can directly create a function which is build with aiges...

• Companion
• cynosure