how to get the pre-trained model "ViT-L-14.tar"

I download the pre-trained model "ViT-L-14.pt"x and its feature is 768. However, the vision_width in yaml file is set 1024. This is different.

Start by locating the YAML file containing vision_width and the code that loads ViT-L-14.pt. Compare the checkpoint's feature dimension with the configured 1024 value and trace which pretrained model artifact the repository expects. Done means the correct model source and the reason for the dimension difference are documented or the mismatch is resolved.

Written by the indexing model from the issue text.