search

TinyLLaVA / TinyLLaVA_Factory

A Framework of Small-scale Large Multimodal Models
https://arxiv.org/abs/2402.14289
Apache License 2.0
658 stars 68 forks source link

Saving vision tower in checkpointing #35

Closed zhangliang-04 closed 8 months ago

zhangliang-04 commented 8 months ago

Dear authors,

Thanks for your great work! I notice that when set tune_entire_model=True, the code will automatically save the trained vision tower model after entire training process. However, it seems that the vision tower is not saved in intermediate checkpoints, which make the checkpoints unusable. Can you provide any suggestions to fix this?

baichuanzhou commented 8 months ago

You can implement saving vision tower in trainer._save method.

zhangliang-04 commented 8 months ago

OK. Thanks for your advice!