search

cambrian-mllm / cambrian

Cambrian-1 is a family of multimodal LLMs with a vision-centric design.
https://cambrian-mllm.github.io/
Apache License 2.0
1.68k stars 110 forks source link

Cannot run offline only with model file #62

Open Rising0321 opened 1 month ago

Rising0321 commented 1 month ago

I download the model checkpoint and want to run cambrian offline. But I encounter the following error:

OSError: We couldn't connect to 'https://huggingface.co' to load this file, couldn't find it in the cached files and it looks like openai/clip-vit-large-patch14-336 is not the path to a directory containing a file named config.json. Checkout your internet connection or see how to run the library in offline mode at 'https://huggingface.co/docs/transformers/installation#offline-mode'.

It seems that this model needs openai/clip-vit-large-patch14-336 to run successfully. Maybe could put it into Readme? Or just put the clip into the model file to avoid download in the offline environment.

Joseph1025 commented 1 month ago

Same here, I tried to change the config.json in cambrian model, but it didn't work. Btw, even if you downloaded the 4 required vision encoders, the way cambrian parse the vision encoders is somehow too complex. It can be really hard to redirect the model to local path.