search

jianzongwu / Language-Driven-Video-Inpainting

(CVPR 2024) Official code for paper "Towards Language-Driven Video Inpainting via Multimodal Large Language Models"
https://arxiv.org/abs/2401.10226
71 stars 3 forks source link

Missing Pre-Trained VAE #8

Closed jishu123456789 closed 5 months ago

jishu123456789 commented 5 months ago

In the paper there was the mention of a Pre-Trained VAE used to extract image features. Where can we find that model.

image

Thank You

jianzongwu commented 5 months ago

Hello, we use the VAE pre-trained from MagicBrush and freeze it during training. You can find the pre-trained VAE here

jishu123456789 commented 5 months ago

Hello sir,

Thank you so much for your prompt response.

Regards Jishu Sen Gupta

On Tue, Jun 11, 2024 at 8:10 AM Jianzong Wu @.***> wrote:

Hello, we use the VAE pre-trained from MagicBrush and freeze it during training. You can find the pre-trained VAE here https://huggingface.co/vinesmsuic/magicbrush-paper

— Reply to this email directly, view it on GitHub https://github.com/jianzongwu/Language-Driven-Video-Inpainting/issues/8#issuecomment-2159662120, or unsubscribe https://github.com/notifications/unsubscribe-auth/A7MUXSOZTZ4RU3RMT3KGFG3ZGZPQNAVCNFSM6AAAAABI4T5FICVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNJZGY3DEMJSGA . You are receiving this because you authored the thread.Message ID: @.*** com>

ᐧ