search

BAAI-DCAI / M3D

M3D: Advancing 3D Medical Image Analysis with Multi-Modal Large Language Models
MIT License
147 stars 7 forks source link

Questions related to M3D-CLIP weights #4

Closed MaochengBai closed 2 months ago

MaochengBai commented 2 months ago

Hello,

Thank you for sharing your project code and associated weights. I would like to inquire about the weights published on Hugging Face under M3D-CLIP/pretrained_ViT.bin. Are these weights extracted from the visual encoder of a CLIP model trained with M3D-Cap? If I want to use the visual encoder of M3D-CLIP, can I directly use the pretrained_ViT.bin weights for this purpose?

Thank you for your time and assistance.

baifanxxx commented 2 months ago

Hi,

Of course, you can use the pretrained_ViT.bin directly because this weight is extracted from the M3D-CLIP.

MaochengBai commented 2 months ago

thank you and may I ask if the dataset includes only CT modalities for medical images?

baifanxxx commented 2 months ago

Hi,

We focus on CT modality now. But the model is universal. You can add other 3D modalities to the dataset directly.

MaochengBai commented 2 months ago

Thank you very much