Jinze Xue

[feat] Option to skip post conda-setup

1

Hi Developers, Thanks for the great work! Currently the post conda-setup step will `remove uncompressed packages to trim down cache folder` in `conda_pkgs_dir` folder , however I have a huge...

【留言板】歌单超 1000 首

248

请留下自己的歌单 URL，博主会尽快回复。

[WIP] CUAEV aev constants memory optimization

1

before: ``` RUN Total AEV Forward Backward Force Optimizer Others Epoch time GPU 0 cu Energy train 21.9 ms 13.1 ms 15.5 ms 0.0 ms 11.0 ms 3.6 ms 65.2...

25% Performance regression from v0.1.1 to 0.2.0 when calculating hessian

40

Hi developers, After I upgraded functorch from `v0.1.1` to `0.2.0`, I noticed a 25% performance regression when calculating hessian, please check the following benchmark result and the attached benchmark script....

vmap does not work with torch.autograd.grad for batched_tensor

3

Related issues: https://github.com/pytorch/functorch/issues/667, https://github.com/pytorch/functorch/issues/704 Hi, I checked the related issues and found that for a batched_tensor, vmap will create a new wrapper and erase the requires_grad filed. So vmap could...

Pin some frequently needed notes on the sidebar with title only, instead of `cmd+F` to search every time. - Sidebar will only shown when screen is large enough - Only...

[使用数据库 (Mongoose)](https://developer.mozilla.org/zh-CN/docs/learn/Server-side/Express_Nodejs/mongoose)

blog2

1

Probably will be abandon.

标签页的标题