-
## 🐛 Bug
There seems to be a discrepancy (in addition to https://github.com/pytorch/xla/issues/3718) in how `torch.nn.Linear` (`torch.nn.functional.linear`) is implemented and dispatched between th…
-
Here I trained a model on all data available from batch 1 in the LINCS dataset, which can be found like this: `aws s3 ls s3://cellpainting-gallery/cpg0004-lincs/broad/workspace/backend/2016_04_01_a549…
-
Relates to [this issue](https://github.com/HamletWantToCode/GPFlux.jl/issues/2).
A MWE an be found in the `examples` directory on [this branch](https://github.com/willtebbutt/Stheno.jl/tree/wct/flu…
-
### 🐛 Describe the bug
When export using torch.jit.trace involves a custom autograd function (activation checkpointing), runtime error occors.
Code snippets (https://github.com/haoheliu/AudioLDM…
-
when I trained, I found that the outcome was not very good, for example the task of opening the drawer turned out to be the opposite, and the training did not complete the task of opening the drawer m…
-
I am really enjoying `jax`, but have found the ergonomics of declaring function arguments as static for use in `jit`, `pjit`, and `pmap` to be a little cumbersome and error-prone when refactoring.
…
-
请问你这边在finetune llama的时候,各个参数如7B,13B,30B各个节点所用资源的数量,有统计过吗?
比方单卡时gpu显存,cpu内存占用?八卡时gpu显存,cpu内存占用?
我现在发现全参数finetune的时候cpu内存占用特别厉害,不知道你这边有没有什么特别的优化没?
-
Hi,
When running the DeepExplain as shown below, I run into the following error. Any suggestions? Please let me know if you need any further info. Thanks!
```from keras import backend as K
impo…
-
The support for js is pretty bad. Most of the issues are in the way of parsing variable names.
Load this example into atom. When I say something is `not tagged` I mean its scope is only `.source.js`…
-
There are some issues with the current Neural Network implementation. Though rusty-machine will not be able to beat existing deep learning frameworks it would be nice if we could provide a simple but …