Open mozhu621 opened 5 months ago
The training data will not be released very soon.
So - essentially - there's no way to reproduce the results? Please correct me if I'm wrong.
So - essentially - there's no way to reproduce the results? Please correct me if I'm wrong.
Currently, you can reproduce the evaluation results using our model on the huggingface hub https://huggingface.co/In2Training/FILM-7B. This repository provides a guide to the evaluation process and our evaluation logs.
If you urgently need to reproduce the training process, you can find training details such as how to construct the training dataset and how to set the training hyperparameters in our preprints https://arxiv.org/abs/2404.16811.
Our data is still in the internal review process and we are trying to move forward. Sorry for any inconvenience.
I didn't find your dataset, and the code about model training with it? Do you make it public?