[项目] 开发一个可以训练的界面

 

 

Nvidia Digits

最后一版 v6.1.1 发布于 2018.04

supported frameworks are: Caffe, Torch, and Tensorflow.

supported platform: Ubuntu 14.04, 16.04 , 也有官方docker image

 

IBM Ffdl

是一个基于K8S 的在分布式硬件上训练的框架，感觉有点重

https://github.com/IBM/FfDL

 

BMW GUI

基于docker技术，可能没法用

https://github.com/BMW-InnovationLab/BMW-TensorFlow-Training-GUI

 

Develop REST API

https://medium.com/nybles/controlling-neural-network-training-from-web-interface-5ef6f0c5bc0d

https://stackoverflow.com/questions/61942505/add-tensorboard-server-to-flask-endpoint

 

 

 

Ref:

用flask 开发rest api, 后端可以接tensorflow, pytorch等， https://medium.com/nybles/controlling-neural-network-training-from-web-interface-5ef6f0c5bc0d

一个现成的用python写的项目， 有GUI，https://github.com/mustafamerttunali/deep-learning-training-gui?tab=readme-ov-file， 受到 Nvidia DIGITS 启发

https://github.com/determined-ai/determined?tab=readme-ov-file, 这个determined 项目好像只看到webui 用来看result 不能用来training

https://aimagazine.com/top10/top-10-no-code-ai-tools

https://www.cbinsights.com/research/no-code-ai-platforms/