leaf-ygq

Results 4 issues of leaf-ygq

Hi @EmilyAlsentzer , I want to know if the code can be run with multiple GPUs. I change the parameter of trainer in train_config.py. ![1](https://user-images.githubusercontent.com/75319189/104708182-52a1b500-5758-11eb-9e7d-25e2903067d9.png) But there was something wrong...

![image](https://github.com/FlagOpen/FlagEmbedding/assets/75319189/a2b24536-cd71-4dd1-8081-dbc30d09ce82) bge-m3dui对显卡gpu要求很高。 之前我用bge-v1.5能够轻轻松松同时对长度为10000+的字符串数组进行编码,但在使用bge-m3同时对长度为10000+的字符串数组进行编码时,报错了,有什么办法能够解决吗?

![image](https://github.com/ztxz16/fastllm/assets/75319189/9c97a0c6-949c-4cce-af4e-f42bc2c83787) 像上图红框框的位置,原来是cpu,代码运行成功,但我想把模型加载在cuda:0卡上,把"cpu"改成"cuda:0"后,报下面这个错误了,要怎么改,才能让模型部署到gpu上呢? ![image](https://github.com/ztxz16/fastllm/assets/75319189/11489a6c-dc46-4ae2-8354-c4b839c62ae8) 另外,也尝试了下面这种方式,模型依旧不能加载到cuda0卡上,大佬指点一下吧 ![image](https://github.com/ztxz16/fastllm/assets/75319189/4b5e2a43-6df2-4461-be35-c303c5c43917)

I have 15,000 pieces of data stored in chromadb. Each piece of data contains (document name, document content). I searched the document names of these 15000 data without modification, but...

2025-review