-
Submitting Author Name: Eunseop Kim
Submitting Author Github Handle: @markean
Repository: https://github.com/markean/melt
Submission type: Pre-submission
Language: en
---
(Copied from s…
-
In the paper [Language models enable zero-shot prediction of the effects of mutations on protein function](https://www.biorxiv.org/content/10.1101/2021.07.09.450648v2) the ESM folks introduce the "Mas…
-
### Feature description
My company is currently working with a development partner for an internal ERP project. I have showcased and done training sessions with 90+ developers to onboard them on us…
-
请教个问题，qwen-7b 输入长度超过12k，推理会内存溢出。
我用了4张卡，模型参数平均分配在4张卡上，但是推理时张量没有分配到4张卡上，只在一张卡上计算，导致内存溢出，分布式推理如何实现？
模型：qwen-1.5-7b-chat
硬件：4张A100-80G
我是这样加载模型的：
self.tokenizer = AutoTokenizer.from_pretrained…
-
Hey some people are still asking about the thermostat are you still working on this?
-
-
Please Help, I am not a coder but I am having issues installing Stable Diffusion. In theory from my understanding at the end I should receive a URL link to go to stablke diffusion but it doesn't appea…
-
你好，我希望用自己的数据集，开始生成embedding以及构建RQVAE跑出来index数据，生成embedding的参考代码是不是缺一些python模块呀，比如load_plm()，我看现在的代码结构并没有这个方法。
-
作者您好！您的工作非常具有启发性！
我们想请教您如何使用BERT-base-uncased进行实验？我们现在将基座PLM改为uncased之后（词表没有改，使用的是您项目中提供的词表），模型似乎不能正常运行。（如下图所示，输出的结果基本上都是0）
![image](https://github.com/JaySaligia/PPT/assets/49402403/ab98d14d-41d9-4…
-
As things are getting hard to maintain I propose a change of overall approach of configuration and execution of TCLB.
**These are notes for a possible direction of development**
- Integrate [TCL…