Jindong Wang comments

Results 60 comments of


                                            Jindong Wang

trafficstars

参与贡献方式 How to contribute

@fzy0728 OK

参与贡献方式 How to contribute

@hwanji 感谢如此细致的指正！已修改！

参与贡献方式 How to contribute

@Xiang-Pan 已添加。这个repo我应该不会再更新了，因为新的内容都在《迁移学习导论》里面了。不过这个repo仍然开放，以前的东西也能用，欢迎你补充更多更好的内容！

迁移学习前可迁移性评估指标

目前没有有效的评估指标。一些已有工作可以以“transferability estimation”为关键词搜索到。但是我的结论是，绝大多数方法做了半天，得到的结论大部分都是我们要选择最大的模型、最大的数据集。因此可能增加模型和数据集的大小才是关键，这也是迁移学习为什么会有效：本身就是要从大数据和大模型上学习知识。

How long was the model trained in total, including the pre-training time before the clip was sent to the clients

@lw0517

RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

Looks like some environment issues. Are you using my docker? Since `root@Oision-Legion-R7000P2021H` doesn't seem to be inside the docker. If so, what kind of GPU and CUDA version of your...

RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

It seems that your GPU is not suitable for training speech tasks. Honestly speaking, speech tasks are really consuming hardware resources and we are doing our experiments on Microsoft Azure...

DIVERSIFY: A General Framework for Time Series Out-of-distribution Detection and Generalization中的cross_dataset的文件

@lw0517 可以看看有的话提供一下

LLaMa 2 inference

Hi @jubick1337 , we are rewriting the code for PromptBench to support more models and make the code easier to use. Check the dev branch: https://github.com/microsoft/promptbench/tree/dev

运行MMD时矩阵不一致问题的报错。

@lw0517 有时间来看一下。