Open cattleyawlotus opened 8 months ago
Thanks for asking! When training and making inference on LLaMA, we use the same environment as ODIE ("On-Demand-Information-Extraction https://arxiv.org/abs/2310.16040) and the packages' versions can be found at https://github.com/yzjiao/On-Demand-IE/blob/9638fb16c29e308aee42b341093d458a00c6c01d/training/requirements.txt. We are have also added the code for the ranking module into the code, so you might also want to update your code as well. Moreover, the pre-training step might take pretty long (a few hours), so we are also uploading the weights to google drive. We will attach a link to our google drive when we are done.
Thanks for asking! When training and making inference on LLaMA, we use the same environment as ODIE ("On-Demand-Information-Extraction https://arxiv.org/abs/2310.16040) and the packages' versions can be found at https://github.com/yzjiao/On-Demand-IE/blob/9638fb16c29e308aee42b341093d458a00c6c01d/training/requirements.txt. We are have also added the code for the ranking module into the code, so you might also want to update your code as well. Moreover, the pre-training step might take pretty long (a few hours), so we are also uploading the weights to google drive. We will attach a link to our google drive when we are done.
Thanks a lot!!!
When I was trying to run the code without changing data and code, I encountered problems about FP16. The problem is the same as #1246 , but it wasn't solved. I suppose I should first examine if there's something wrong with the environment. Thanks.