-
Thank you for your released code and weights!I see that in the paper, your idea is to start from LISA, I wonder if you have reproduced the training phase of LISA, and whether you have reached the perf…
-
@lisamr Has requested the 2020-2023 V3 Feds data as it is generated. Lisa is interested in CA for benchmarking reasons. Would be useful to connect to the EIS bucket. Already generated the data, but w…
-
In the article, only the comparison of the average weight paradigm of each layer during lora fine-tuning is given.
1. But what if their weights are different before fine-tuning?
2. Using the weighte…
-
-
Has anyone managed to replicate the outcomes of lisa-7b? I'm puzzled as to why, after ten epochs of training, my test results for giou are 0.05 and for ciou are 0.03.
-
Is there any Lisa that supports llama3.1 yet?
-
При загрузки файлов на проверку в lisa происходит бесконечная загрузка, узнал о проблеме через браузер.
Приложение было установлено согласно инструкциям github в docker compose.
OS - centos 8
Сбор…
-
https://drive.google.com/file/d/1gFtPT4Ekd98H-oc_Mr6VgWi9T0LFYLNh/view?usp=sharing
This app broke since 247d135f67dde9df0d32edd3315385ad8e7dc962 and further
**st/nine: Implement SYSTEMMEM buffers sa…
-
hi!
could any one konw how to get the LISA's repository?the prior version of LISA2?Thanks!
-
**Describe the bug**
I think there might be something wrong with the current LISA implementation. There is no difference in training loss, no matter how many layers are active.
Not using LMFlow bu…