Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Is the linear layer initialized by llava's linear layer? #124

Open
vhzy opened this issue Aug 15, 2024 · 1 comment
Open

Is the linear layer initialized by llava's linear layer? #124

vhzy opened this issue Aug 15, 2024 · 1 comment

Comments

@vhzy
Copy link

vhzy commented Aug 15, 2024

Is the linear layer initialized by the linear layer of llava? I found that the pretrain_mm_mlp_adapter parameter is not set in the script. Does it mean that the linear layer is not initialized by llava?

@ChangPtR
Copy link

@vhzy hello, I get the same question. According to the paper "We use LLaVA (Liu et al., 2023) as our baseline model and finetune it on our 100K video instruction pairs. We only update the linear layer projecting the video features to the LLMs’ input space, while the rest of the architecture is kept frozen", the linear layer should be initialized by llava, but there is no mm_projector.bin in LLaVA-7B-Lightening-v1-1.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants