Li Mu Zhi

Results 7 issues of Li Mu Zhi

In README file, Kubeflow Pipelines use docker or pns executor as argo workflow executor. according document of argo workflow.https://argoproj.github.io/argo-workflows/workflow-executors recommends Emissary Executor. I also find configuration files in apps/pipeline/upstream/env/platform-agnostic-multi-user-emissary. so...

您好,我在执行`python evaluate_models.py --model chatglm`想要复现你在README里面的评估结果都时候发现执行的结果好像不太一一致,差距还挺大,请问是什么原因呢? 我的执行结果如下 Sentence Level: acc:0.2627, precision:0.2472, recall:0.2413, f1:0.2442, cost time:131.01 s, total num: 1100 看代码里面用的模型是chatglm3-6b和一个您训练的lora,都是用的库里的,没有修改

bug

if we have notes as tree: - inbox - first.md - second.md - journal - first.md - second.md if set images root path in /assets, hope to have same directory...

seq2seq_model在输入nezha模型之前的tokenizer设置了最大的长度1024,请问一下这个是什么考虑。感觉很多文本都被截断了。

When I check code about Chat Streaming, This variable seems to be shared by multiple processes, I also find this variable is assigned by another process. But `is_end` is local...

enhancement

starts with "./" use relative path, otherwise in root of obsidian vault root.

### Describe your problem I noticed that es was used as a hybrid search solution (combining keyword search and vector search) in the project. In a simple test case, I...

question