conda安装di-engine，无法与python 3.9兼容

[Rocky-CN]

更新：

pip可以正常安装

情况描述;

Pytorch 2.x 至少需要python 3.8+，所以conda创建环境时选择了python3.9
但是当尝试通过conda命令安装di-engine，失败了。
conda命令如下：

conda install -c opendilab di-engine

错误日志如下：

Platform: linux-64
Collecting package metadata (repodata.json): done
Solving environment: failed

LibMambaUnsatisfiableError: Encountered problems while solving:
  - nothing provides python_abi 3.6.* *_cp36m needed by di-engine-v0.2.2-py36_0

Could not solve for environment specs
The following packages are incompatible
└─ di-engine is not installable because there are no viable options
   ├─ di-engine [v0.2.2|v0.3.0|v0.3.1|v0.4.0] would require
   │  └─ python_abi 3.6.* *_cp36m, which does not exist (perhaps a missing channel);
   └─ di-engine [v0.4.1|v0.4.2|...|v0.5.2] would require
      └─ python_abi 3.7.* *_cp37m, which does not exist (perhaps a missing channel).

[PaParaZz1]

可能是之前 conda 的 release 版本有点问题，我们会检查修复一下，下个版本会放宽 Python 版本限制为3.7-3.10

[Rocky-CN]

pip安装后，检查 DI-engine 是否可用时，也会有两个warning和一个error
命令行如下：

ding -v

日志如下：

/usr/local/lib/python3.10/dist-packages/treevalue/tree/integration/torch.py:23: FutureWarning: `torch.utils._pytree._register_pytree_node` is deprecated. Please use `torch.utils._pytree.register_pytree_node` instead.
  register_for_torch(TreeValue)
/usr/local/lib/python3.10/dist-packages/treevalue/tree/integration/torch.py:24: FutureWarning: `torch.utils._pytree._register_pytree_node` is deprecated. Please use `torch.utils._pytree.register_pytree_node` instead.
  register_for_torch(FastTreeValue)
ERROR:root:Please install pyecharts first, you can install it by running 'pip install pyecharts'

虽然在执行pip install pyecharts后，问题消失。
但希望也能修复下，毕竟对于新手来说，跟着教程安装在第一步就出现一堆错误，挺劝退的...

[PaParaZz1]

感谢反馈，我们已经将 pyecharts 相关的这个 log 从error将为warning，这个是用在显存分析工具（link）里进行可视化的库，是常规训练流程所不必须的，不影响一般训练任务，所以没有放在setup.py里。