search

DAMO-NLP-SG / Video-LLaMA

[EMNLP 2023 Demo] Video-LLaMA: An Instruction-tuned Audio-Visual Language Model for Video Understanding
BSD 3-Clause "New" or "Revised" License
2.77k stars 255 forks source link

The question about llama parameters during pre-training and fine-tuning. #130

Closed cooper12121 closed 10 months ago

cooper12121 commented 10 months ago

"During pre-training and fine-tuning, are the parameters of the llama frozen?"

lixin4ever commented 10 months ago

Right, the parameters of LLMs are kept frozen during both pre-training and fine-tuning.

cooper12121 commented 10 months ago

Right, the parameters of LLMs are kept frozen during both pre-training and fine-tuning.

Translate: Thank you for your reply. Have you tested some multimodal tasks such as multimodal NER and relation extraction? How does only fine-tuning Q-former perform when llama parameters are frozen?